3. Mechanism and risk factors of neurologic injury
The spectrum of post-operative neurologic injury can be schematically
outlined as:
- intra-operative and early post-operative events – deficit diagnosed
on awakening at extubation, mainly related to atheroembolism from
aortic manipulation or cerebral hypoperfusion;
- delayed post-operative events – occurring within 30 post-operative
days, but after a symptom-free interval following extubation (mainly
secondary to atrial fibrillation, low cardiac output states, bleeding
and hypoperfusion);
- late events – occurring after the 30-day post-operative period.
Coronary artery bypass performed on an arrested heart relies on aortic
cannulation and cross-clamping; both steps represent considerable aortic
manipulation associated with significant atheroembolic risk. The
mechanisms for this to occur include detachment of atherosclerotic
material from the aortic intima as a result of external manipulation
(cannulation, aortic cross-clamping and partial occlusion); or internal
disruption caused by the ”sandblasting” effect of the high-velocity jet
of blood exiting the aortic cannula and impacting on the atherosclerotic
intima or pedunculated atheroma.
The role of the “porcelain aorta” (43) (Figure 1) in the pathogenesis
of adverse neurologic outcome after isolated CABGs has been postulated
for almost four decades (3-6).
Multiple cadaveric studies have documented a high embolic burden to the
brain during CPB and aortic manipulation in the presence of ascending
aortic atherosclerosis (7,8). This is manifested by specific cerebral
microvascular alterations in the form of small capillary and arteriolar
dilatation (7). In an autopsy study by Blauth (8) 21.7% of patients who
died after cardiac surgery on CPB had evidence of atheroemboli (or
abnormalities consistent with atheroemboli) in the brain; severe aortic
atherosclerosis and coronary surgery were significantly associated with
higher risk of neurologic embolization.
In a multicenter prospective study that enrolled 2,108 patients
undergoing elective CABG in 24 U.S. medical institutions, Roach (9)
reported an overall incidence of type I cerebral events (coma, stroke or
TIA) of 3.1%; moderate to severe aortic atherosclerosis assessed by
manual palpation of the ascending aorta was the main risk factor,
increasing the rate of postoperative stroke four-fold. Similar findings
were reported by Stamou (10) on a series of 16,528 patients who
underwent isolated on-pump CABG with an overall stroke rate of 2.0%;
atherosclerosis of the ascending aorta and prolonged bypass and
cross-clamp time were significantly more frequent in patients with
postoperative neurologic complications. Data from Tarakji (1) showed an
overall stroke rate of 1.6% in a large series of consecutive isolated
CABG performed over a period of 30 years. The main risk factors were
older age and variables representing higher atherosclerotic burden like
PVD and carotid stenosis. Minimization of aortic manipulation with
off-pump or on-pump beating heart techniques provided the lowest risk.
In a review of 6,682 consecutive patients undergoing isolated on-pump
CABG performed with single aortic clamp technique and without epiaortic
scanning, Borger (11) reported a prevalence of postoperative stroke of
1.5%, associated with a significant increase of in-hospital mortality
(p<0.001). The atheroembolic etiology of those strokes was
supported by radiological and pathological evidence of ischemic lesions
in the territory distribution of major cerebral vessels; surgical
manipulation (cannulation, cross-clamping, proximal anastomosis) of an
atherosclerotic ascending aorta was the most common embolic source.
Aortic calcifications were once again found to increase threefold the
risk of post-operative stroke (overall rate 1.4%) in a review of 19,224
patients undergoing isolated on-pump CABG in 31 hospitals in the State
of New York by John (12). As for other studies, stroke patients
experienced a significantly increased in-hospital mortality (24.8% vs
2.0%, p<0.001). Likewise, Rao (14) reported a stroke rate of
1.4% among a series of 3,910 consecutive isolated on-pump CABGs (single
clamp technique). Severe coronary artery disease was found to be the
most important predictor of stroke, followed by older age, previous
TIA/stroke, PVD and diabetes: all those factors clearly represent a
surrogate for aortic atherosclerosis. A higher stroke rate of 4.3% was
reported by Lynn (15) in a series of 1,000 consecutive on-pump CABGs;
extensive aortic calcifications noted by the surgeon was a highly
significant risk factor for permanent neurologic deficit (11 strokes in
57 patients with aortic calcification, 19%; p<0.0001).
Finally, the investigators of the Multicenter Study of Perioperative
Ischemia (13) developed a preoperative stroke risk index used to
estimate the individual risk of postoperative stroke. Advanced age, PVD
and diabetes were strong predictors of atherosclerotic disease of the
aorta and cerebral vessels, hence the authors implicated embolization as
the likely primary aetiology of major neurologic deficits after CABG.
Early postoperative strokes are more often right-sided compared with
delayed strokes that are more uniformly distributed. Data from Boivie
and Hedberg (44-46) showed clinical and radiological evidence that
right-sided perioperative events were largely preponderant compared to
contralateral strokes. Manipulation of a calcified ascending aorta
explains the embolic nature of particles tangentially expelled into the
brachiocephalic artery.
Grooters (47) looked at the effect of perfusion jets and
“sandblasting” from CPB aortic cannulae into the aortic arch. Using
transesophageal echo (TEE) they demonstrated that both end-hole and
side-holes cannulae generated high velocity jets directed against the
aortic wall, potentially responsible for the disruption of friable
atherosclerotic plaques. On the other hand, the Dispersion cannula
showed a broad wedge-like perfusion pattern with significantly lower
exit velocities.
The effect of aortic clamping has also been extensively investigated. In
a cadaveric study from Boivie (48), ten calcified ascending aortas were
mounted on a perfusion model, clamped multiple times and finally washed
out. The investigators found that a significant amount of particular
matter had been released, comprised of both calcified and cellular
material (average diameter of 0.63 ± 0.03 mm). Although the majority of
the embolic load occurred following the first clamping, subsequent
clamps still continued to release particulate matter and therefore this
questions multiple-clamp surgical techniques (Figure 2). Barbut (49)
looked at the embolic burden detected by transcranial doppler (TCD)
during on-pump CABG using a double-clamping technique. The largest
number of embolic signals (58%) were noted at aortic cross-clamp or
partial-occlusion clamp removal (seven-fold and five-fold higher rates
respectively) especially when higher grade of aortic atheroma were
present (p<0.05).
Watters (50) detected a significant decrease in TCD embolic signals in
patients undergoing off-pump CABG (OPCAB) with partial occlusion clamp,
compared to on-pump CABG with double clamping (median embolic signals 79vs 3 respectively). Although this study was non-randomized, the
two groups were similar in terms of Parsonnet risk score, number of
grafts performed and overall presence of aortic calcification. Finally,
in another study involving TCD monitoring during CABG, Taylor (51)
showed that a relevant number of embolic events while on CPB were
recorded during perfusionist interventions (drugs administration and
blood sampling) in the form of small air bubbles.