Portal de Periódicos da CAPES

Perioperative neuropsychologic testing

2001; Elsevier BV; Volume: 15; Issue: 1 Linguagem: Inglês

10.1053/jcan.2001.2028

1532-8422

John M. Murkin,

Intensive Care Unit Cognitive Disorders

THE PROBLEM of perioperative brain injury in cardiac surgical patients is of paramount importance. This issue contains 3 original investigations and 1 methodologic review concerned with perioperative neuropsychologic testing in cardiac surgical patients. Because of this interesting juxtaposition, it is recommended that the article on methodologic considerations by Slade et al be read before the articles by Andrew et al, Silbert et al, and Ebert et al.The use of cognitive testing has proved to be of immense value in bringing increased recognition to the problem of perioperative brain injury as well as in providing a sensitive means of assessing overall outcomes (see the article by Andrew et al) and in detecting beneficial results associated with specific interventions.1Patel RL Turtle MR Chambers DJ et al.Alpha-stat acid-base regulation during cardiopulmonary bypass improves neuropsychologic outcome in patients undergoing coronary artery bypass grafting.J Thorac Cardiovasc Surg. 1996; 111: 1267-1279Abstract Full Text Full Text PDF PubMed Scopus (117) Google Scholar, 2Murkin JM Martzke JS Buchan AM et al.A randomized study of the influence of perfusion technique and pH management strategy in 316 patients undergoing coronary artery bypass surgery: Part 2. Neurological and cognitive outcomes.J Thorac Cardiovasc Surg. 1995; 110: 349-362Abstract Full Text Full Text PDF PubMed Scopus (250) Google Scholar Caution must be exercised in interpreting the results derived from such studies, however. The mere ability to do a thing is not sufficient alone to justify its performance. With regards to cognitive testing in the early postoperative period, having a patient group capable of producing a score on (some) cognitive tests does not imply de facto demonstration of the successful use of early cognitive testing (see the article by Silbert et al). As discussed by Slade et al, many important considerations must be addressed before such data can be meaningfully interpreted.For their study of early postoperative cognitive testing, Silbert et al state that they investigated 50 patients. These 50 patients provided the baseline data the authors used for determining dysfunction (ie, a decline in change score [preoperative − postoperative score] of magnitude >1 SD of the average baseline score for that test). The immediate problem that arises is due to the fact that in reality only “30 patients completed ≥4 tests on both occasions.” The sample size in Silbert's study is at most 30, not 50, and the inclusion of the 20 patients who did not complete the study skews the data. It is most likely that the 20 patients who were unable to start or to complete the postoperative test battery were older, were sicker, and had greater comorbidity, all factors that would tend to significantly decrease their performance during preoperative testing. This inclusion would result in a larger SD for the baseline test score data, giving rise to significantly less likelihood of detecting dysfunction postoperatively (ie, a patient's score would have to drop that much further as the baseline SD increases before meeting the threshold defining dysfunction). This situation is precisely what Silbert et al observed when reporting an incidence of dysfunction of 10% at 5 days postoperatively, in contradistinction to the incidences of 30% to 50% as reported in many other studies1Patel RL Turtle MR Chambers DJ et al.Alpha-stat acid-base regulation during cardiopulmonary bypass improves neuropsychologic outcome in patients undergoing coronary artery bypass grafting.J Thorac Cardiovasc Surg. 1996; 111: 1267-1279Abstract Full Text Full Text PDF PubMed Scopus (117) Google Scholar, 2Murkin JM Martzke JS Buchan AM et al.A randomized study of the influence of perfusion technique and pH management strategy in 316 patients undergoing coronary artery bypass surgery: Part 2. Neurological and cognitive outcomes.J Thorac Cardiovasc Surg. 1995; 110: 349-362Abstract Full Text Full Text PDF PubMed Scopus (250) Google Scholar (also see Andrew et al).Further compounding this problem is the fact that despite the allusion to the Consensus Statement,3Murkin JM Newman S Stump DA et al.Statement of consensus on assessment of neurobehavioral outcomes after cardiac surgery.Ann Thorac Surg. 1995; 59: 1289-1295Abstract Full Text PDF PubMed Scopus (526) Google Scholar of the 4 necessary but not sufficient recommended tests (Rey auditory verbal learning test, Trail Making A and B, and grooved pegboard) only 2, Trail Making parts A and B, were (possibly) used by Silbert et al. Having then initially defined a battery of 6 tests, the fact that “the complete battery of neuropsychological tests and visual analog scales were completed by 16 (32%) patients” introduces a further confound. According to the original study protocol, only 32% of patients were able to complete these evaluations. That fact should be their conclusion and should have been the focus of their discussion.To summarily change the study design to include patients who completed only ≥4 neuropsychologic tests and yet still define dysfunction as a decrease in ≥2 individual tests (ie, requiring a significant decline in 50% of tests administered) acts to compound the error. This error would further bias the data against detecting cognitive dysfunction despite its having been present (ie, a type II error). Confounds introduced by practice effects are not even considered.As discussed by Slade et al, practice effects are a function of specific processes as well as the specific content involved with cognitive testing. Substituting memory subtests in lieu of those recommended,3Murkin JM Newman S Stump DA et al.Statement of consensus on assessment of neurobehavioral outcomes after cardiac surgery.Ann Thorac Surg. 1995; 59: 1289-1295Abstract Full Text PDF PubMed Scopus (526) Google Scholar with the recognition that “lack of alternate forms exposes patients to a practice effect,” because “the utility of the tests far outweighed the problem of practice effects, which in this setting would have skewed the results in favor of not detecting a change,” then commenting that the apparently lower incidence of cognitive dysfunction seen at 5 days in contradistinction to higher incidences reported elsewhere “raises doubts” about the utility of later testing as an endpoint (see Silbert et al) reflects some profound misunderstanding of these subtle but fundamental issues. Hence, the necessity of a methodologic review such as that by Slade et al.Andrew et al take a much different approach. For their study comparing neuropsychologic outcomes in patients after heart valve or coronary artery bypass graft (CABG) surgery, they employed a control group matched for age, gender, and education. Assessment of such a control group was one of the recommendations of the Consensus Statement,3Murkin JM Newman S Stump DA et al.Statement of consensus on assessment of neurobehavioral outcomes after cardiac surgery.Ann Thorac Surg. 1995; 59: 1289-1295Abstract Full Text PDF PubMed Scopus (526) Google Scholar an approach that has been incorporated successfully into several previous such studies.2Murkin JM Martzke JS Buchan AM et al.A randomized study of the influence of perfusion technique and pH management strategy in 316 patients undergoing coronary artery bypass surgery: Part 2. Neurological and cognitive outcomes.J Thorac Cardiovasc Surg. 1995; 110: 349-362Abstract Full Text Full Text PDF PubMed Scopus (250) Google Scholar, 4Kneebone AC Andrew MJ Baker RA Knight JL Neuropsychological change following cardiac surgery: Use of reliable change indices.Ann Thorac Surg. 1998; 65: 1320-1325Abstract Full Text Full Text PDF PubMed Scopus (82) Google ScholarAs discussed by Slade et al, use of an appropriate control group allows the magnitude of the practice effect to be estimated and corrected for. Kneebone et al4Kneebone AC Andrew MJ Baker RA Knight JL Neuropsychological change following cardiac surgery: Use of reliable change indices.Ann Thorac Surg. 1998; 65: 1320-1325Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar showed the utility of the reliable change index, which can be derived from control group data and enables correction for practice effects as well as provides a more compelling rationale for the selection of threshold scores for defining cognitive dysfunction. The use of the reliable change index represents a rational and powerful way of standardizing the approach to the study of perioperative cognitive dysfunction, and it is hoped that it will be more widely adopted.Based on this approach, analysis of the data of Andrew et al has yielded the interesting result that intermediate-term (6 months) outcomes differ between patients undergoing heart valve versus CABG surgery, with greater recovery being shown in the CABG group. Because it is well established that cerebral embolization is exponentially greater in valve versus CABG procedures, it appears increasingly apparent that such events, likely reflecting a preponderance of microgaseous air bubbles, are not innocuous and may have longer term sequelae. Further circumstantial evidence of the detrimental role of cerebral air emboli can be derived from the study of Borger et al,5Borger MA, Peniston CM, Weisel RD, et al: Neuropsychological impairment post-coronary bypass surgery: Effect of gaseous microemboli during perfusionist interventions. J Thorac Cardiovasc Surg (in press)Google Scholar in which they showed a positive correlation among number of perfusionist interventions (drug injection, blood aspiration into and from the bypass circuit), cerebral embolization, and postoperative cognitive dysfunction. The issue of microgaseous air emboli deserves further study.Ebert et al take a slightly different approach. They compare cognitive performance and neurologic outcomes (eg, neurobehavioral performance) between CABG and valve surgery patients in the early postoperative period. In this case, however, the assessment intervals averaged 2.5 days and 7 days postoperatively. For their definition of dysfunction, Ebert et al used a 1.5 SD change in score on ≥2 tests, of what was apparently a test battery of 5 cognitive tests (ie, a significant decrease in score on ≥40% of tests administered). Ebert et al also employed a large sample of patients undergoing elective cardiac surgery to normalize their group scores for age, education, and other variables so that the 2 groups were comparable from a preoperative cognitive performance basis. It would have been especially informative had Ebert et al calculated reliable change indices from their control patients as the basis for defining dysfunction. In their study, the absolute values for incidences of dysfunction seen in the groups (ie, 71% and 57% at 2.5 days and 36% and 19% at 7 days for valve and CABG groups) fall roughly midway between those of Silbert et al and those of Andrew et al and others.1Patel RL Turtle MR Chambers DJ et al.Alpha-stat acid-base regulation during cardiopulmonary bypass improves neuropsychologic outcome in patients undergoing coronary artery bypass grafting.J Thorac Cardiovasc Surg. 1996; 111: 1267-1279Abstract Full Text Full Text PDF PubMed Scopus (117) Google Scholar, 2Murkin JM Martzke JS Buchan AM et al.A randomized study of the influence of perfusion technique and pH management strategy in 316 patients undergoing coronary artery bypass surgery: Part 2. Neurological and cognitive outcomes.J Thorac Cardiovasc Surg. 1995; 110: 349-362Abstract Full Text Full Text PDF PubMed Scopus (250) Google Scholar It is also likely that through the greater sensitivity introduced by standardizing preoperative cognitive performance between valve and CABG groups, Ebert et al were able to differentiate a significant difference in outcomes, even at these early postoperative intervals. This situation is in contradistinction to the findings of Andrew et al, in which the 6-month follow-up, rather than the 7-day assessment, showed decline in the valve group relative to the CABG group.The use of pH-stat management by Ebert et al for valve and CABG patients during cardiopulmonary bypass is of interest because an overall decrease in cognitive performance in patients subjected to pH-stat versus α-stat management during cardiopulmonary bypass has been convincingly shown in 2 separate clinical trials employing >400 CABG surgery patients.1Patel RL Turtle MR Chambers DJ et al.Alpha-stat acid-base regulation during cardiopulmonary bypass improves neuropsychologic outcome in patients undergoing coronary artery bypass grafting.J Thorac Cardiovasc Surg. 1996; 111: 1267-1279Abstract Full Text Full Text PDF PubMed Scopus (117) Google Scholar, 2Murkin JM Martzke JS Buchan AM et al.A randomized study of the influence of perfusion technique and pH management strategy in 316 patients undergoing coronary artery bypass surgery: Part 2. Neurological and cognitive outcomes.J Thorac Cardiovasc Surg. 1995; 110: 349-362Abstract Full Text Full Text PDF PubMed Scopus (250) Google Scholar This higher rate of injury has been shown to be due to the significantly higher numbers of cerebral emboli associated with pH-stat management, a consequence of the attendant cerebral hyperperfusion.1Patel RL Turtle MR Chambers DJ et al.Alpha-stat acid-base regulation during cardiopulmonary bypass improves neuropsychologic outcome in patients undergoing coronary artery bypass grafting.J Thorac Cardiovasc Surg. 1996; 111: 1267-1279Abstract Full Text Full Text PDF PubMed Scopus (117) Google Scholar Whether the use of pH-stat management had a relatively greater effect on the valve group in this study, given the exponentially higher rates of cerebral embolization associated with heart valve procedures, is unclear.Ebert et al postulate an anatomic basis for the observed cognitive dysfunction (ie, memory performance) imputing selective hippocampal vulnerability to hypoxia as a possible mechanism. In considering this possibility, Ebert et al are well in line with the thinking of clinicians who consider the evidence for cerebral microemboli, particularly as associated with use of cardiotomy suction,6Brooker JF Brown WR Moody DM et al.Cardiotomy suction: A major source of brain lipid emboli during cardiopulmonary bypass.Ann Thorac Surg. 1998; 65: 1651-1655Abstract Full Text Full Text PDF PubMed Scopus (221) Google Scholar to be consistent with this observation. It appears most likely that microemboli-induced regional cerebral ischemia, with secondary regional hypoxia, is at root cause, rather than systemic hypoxia per se. Given that patients undergoing heart valve surgery suffer exponentially more emboli than do patients undergoing CABG surgery, predominantly of a microgaseous nature, this strongly imputes a significant negative outcome to microgaseous air embolization.In summary, the efforts of all these groups are to be applauded. The time, effort, and degree of commitment necessary to undertake any such clinical studies in this complex area are huge. Andrew et al continue to establish and employ important new methodologic standards for this line of investigation. Their study and that of Ebert et al are further evidence for the putative role of microgaseous air in the genesis of some forms of postoperative brain injury. Silbert et al are to be encouraged in their efforts. Early postoperative cognitive performance is a potentially important and sensitive discriminator to evaluate new techniques and technologies. Given the experiences of Silbert et al, perhaps one meaningful endpoint in the early postoperative period is actual numbers of patients able to successfully complete such postoperative cognitive testing per se, as analyzed for anesthetic technique or other types of interventions. Attention to methodologic details, and incorporation of control group-derived reliable change indices can only enhance further understanding of these issues. THE PROBLEM of perioperative brain injury in cardiac surgical patients is of paramount importance. This issue contains 3 original investigations and 1 methodologic review concerned with perioperative neuropsychologic testing in cardiac surgical patients. Because of this interesting juxtaposition, it is recommended that the article on methodologic considerations by Slade et al be read before the articles by Andrew et al, Silbert et al, and Ebert et al. The use of cognitive testing has proved to be of immense value in bringing increased recognition to the problem of perioperative brain injury as well as in providing a sensitive means of assessing overall outcomes (see the article by Andrew et al) and in detecting beneficial results associated with specific interventions.1Patel RL Turtle MR Chambers DJ et al.Alpha-stat acid-base regulation during cardiopulmonary bypass improves neuropsychologic outcome in patients undergoing coronary artery bypass grafting.J Thorac Cardiovasc Surg. 1996; 111: 1267-1279Abstract Full Text Full Text PDF PubMed Scopus (117) Google Scholar, 2Murkin JM Martzke JS Buchan AM et al.A randomized study of the influence of perfusion technique and pH management strategy in 316 patients undergoing coronary artery bypass surgery: Part 2. Neurological and cognitive outcomes.J Thorac Cardiovasc Surg. 1995; 110: 349-362Abstract Full Text Full Text PDF PubMed Scopus (250) Google Scholar Caution must be exercised in interpreting the results derived from such studies, however. The mere ability to do a thing is not sufficient alone to justify its performance. With regards to cognitive testing in the early postoperative period, having a patient group capable of producing a score on (some) cognitive tests does not imply de facto demonstration of the successful use of early cognitive testing (see the article by Silbert et al). As discussed by Slade et al, many important considerations must be addressed before such data can be meaningfully interpreted. For their study of early postoperative cognitive testing, Silbert et al state that they investigated 50 patients. These 50 patients provided the baseline data the authors used for determining dysfunction (ie, a decline in change score [preoperative − postoperative score] of magnitude >1 SD of the average baseline score for that test). The immediate problem that arises is due to the fact that in reality only “30 patients completed ≥4 tests on both occasions.” The sample size in Silbert's study is at most 30, not 50, and the inclusion of the 20 patients who did not complete the study skews the data. It is most likely that the 20 patients who were unable to start or to complete the postoperative test battery were older, were sicker, and had greater comorbidity, all factors that would tend to significantly decrease their performance during preoperative testing. This inclusion would result in a larger SD for the baseline test score data, giving rise to significantly less likelihood of detecting dysfunction postoperatively (ie, a patient's score would have to drop that much further as the baseline SD increases before meeting the threshold defining dysfunction). This situation is precisely what Silbert et al observed when reporting an incidence of dysfunction of 10% at 5 days postoperatively, in contradistinction to the incidences of 30% to 50% as reported in many other studies1Patel RL Turtle MR Chambers DJ et al.Alpha-stat acid-base regulation during cardiopulmonary bypass improves neuropsychologic outcome in patients undergoing coronary artery bypass grafting.J Thorac Cardiovasc Surg. 1996; 111: 1267-1279Abstract Full Text Full Text PDF PubMed Scopus (117) Google Scholar, 2Murkin JM Martzke JS Buchan AM et al.A randomized study of the influence of perfusion technique and pH management strategy in 316 patients undergoing coronary artery bypass surgery: Part 2. Neurological and cognitive outcomes.J Thorac Cardiovasc Surg. 1995; 110: 349-362Abstract Full Text Full Text PDF PubMed Scopus (250) Google Scholar (also see Andrew et al). Further compounding this problem is the fact that despite the allusion to the Consensus Statement,3Murkin JM Newman S Stump DA et al.Statement of consensus on assessment of neurobehavioral outcomes after cardiac surgery.Ann Thorac Surg. 1995; 59: 1289-1295Abstract Full Text PDF PubMed Scopus (526) Google Scholar of the 4 necessary but not sufficient recommended tests (Rey auditory verbal learning test, Trail Making A and B, and grooved pegboard) only 2, Trail Making parts A and B, were (possibly) used by Silbert et al. Having then initially defined a battery of 6 tests, the fact that “the complete battery of neuropsychological tests and visual analog scales were completed by 16 (32%) patients” introduces a further confound. According to the original study protocol, only 32% of patients were able to complete these evaluations. That fact should be their conclusion and should have been the focus of their discussion. To summarily change the study design to include patients who completed only ≥4 neuropsychologic tests and yet still define dysfunction as a decrease in ≥2 individual tests (ie, requiring a significant decline in 50% of tests administered) acts to compound the error. This error would further bias the data against detecting cognitive dysfunction despite its having been present (ie, a type II error). Confounds introduced by practice effects are not even considered. As discussed by Slade et al, practice effects are a function of specific processes as well as the specific content involved with cognitive testing. Substituting memory subtests in lieu of those recommended,3Murkin JM Newman S Stump DA et al.Statement of consensus on assessment of neurobehavioral outcomes after cardiac surgery.Ann Thorac Surg. 1995; 59: 1289-1295Abstract Full Text PDF PubMed Scopus (526) Google Scholar with the recognition that “lack of alternate forms exposes patients to a practice effect,” because “the utility of the tests far outweighed the problem of practice effects, which in this setting would have skewed the results in favor of not detecting a change,” then commenting that the apparently lower incidence of cognitive dysfunction seen at 5 days in contradistinction to higher incidences reported elsewhere “raises doubts” about the utility of later testing as an endpoint (see Silbert et al) reflects some profound misunderstanding of these subtle but fundamental issues. Hence, the necessity of a methodologic review such as that by Slade et al. Andrew et al take a much different approach. For their study comparing neuropsychologic outcomes in patients after heart valve or coronary artery bypass graft (CABG) surgery, they employed a control group matched for age, gender, and education. Assessment of such a control group was one of the recommendations of the Consensus Statement,3Murkin JM Newman S Stump DA et al.Statement of consensus on assessment of neurobehavioral outcomes after cardiac surgery.Ann Thorac Surg. 1995; 59: 1289-1295Abstract Full Text PDF PubMed Scopus (526) Google Scholar an approach that has been incorporated successfully into several previous such studies.2Murkin JM Martzke JS Buchan AM et al.A randomized study of the influence of perfusion technique and pH management strategy in 316 patients undergoing coronary artery bypass surgery: Part 2. Neurological and cognitive outcomes.J Thorac Cardiovasc Surg. 1995; 110: 349-362Abstract Full Text Full Text PDF PubMed Scopus (250) Google Scholar, 4Kneebone AC Andrew MJ Baker RA Knight JL Neuropsychological change following cardiac surgery: Use of reliable change indices.Ann Thorac Surg. 1998; 65: 1320-1325Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar As discussed by Slade et al, use of an appropriate control group allows the magnitude of the practice effect to be estimated and corrected for. Kneebone et al4Kneebone AC Andrew MJ Baker RA Knight JL Neuropsychological change following cardiac surgery: Use of reliable change indices.Ann Thorac Surg. 1998; 65: 1320-1325Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar showed the utility of the reliable change index, which can be derived from control group data and enables correction for practice effects as well as provides a more compelling rationale for the selection of threshold scores for defining cognitive dysfunction. The use of the reliable change index represents a rational and powerful way of standardizing the approach to the study of perioperative cognitive dysfunction, and it is hoped that it will be more widely adopted. Based on this approach, analysis of the data of Andrew et al has yielded the interesting result that intermediate-term (6 months) outcomes differ between patients undergoing heart valve versus CABG surgery, with greater recovery being shown in the CABG group. Because it is well established that cerebral embolization is exponentially greater in valve versus CABG procedures, it appears increasingly apparent that such events, likely reflecting a preponderance of microgaseous air bubbles, are not innocuous and may have longer term sequelae. Further circumstantial evidence of the detrimental role of cerebral air emboli can be derived from the study of Borger et al,5Borger MA, Peniston CM, Weisel RD, et al: Neuropsychological impairment post-coronary bypass surgery: Effect of gaseous microemboli during perfusionist interventions. J Thorac Cardiovasc Surg (in press)Google Scholar in which they showed a positive correlation among number of perfusionist interventions (drug injection, blood aspiration into and from the bypass circuit), cerebral embolization, and postoperative cognitive dysfunction. The issue of microgaseous air emboli deserves further study. Ebert et al take a slightly different approach. They compare cognitive performance and neurologic outcomes (eg, neurobehavioral performance) between CABG and valve surgery patients in the early postoperative period. In this case, however, the assessment intervals averaged 2.5 days and 7 days postoperatively. For their definition of dysfunction, Ebert et al used a 1.5 SD change in score on ≥2 tests, of what was apparently a test battery of 5 cognitive tests (ie, a significant decrease in score on ≥40% of tests administered). Ebert et al also employed a large sample of patients undergoing elective cardiac surgery to normalize their group scores for age, education, and other variables so that the 2 groups were comparable from a preoperative cognitive performance basis. It would have been especially informative had Ebert et al calculated reliable change indices from their control patients as the basis for defining dysfunction. In their study, the absolute values for incidences of dysfunction seen in the groups (ie, 71% and 57% at 2.5 days and 36% and 19% at 7 days for valve and CABG groups) fall roughly midway between those of Silbert et al and those of Andrew et al and others.1Patel RL Turtle MR Chambers DJ et al.Alpha-stat acid-base regulation during cardiopulmonary bypass improves neuropsychologic outcome in patients undergoing coronary artery bypass grafting.J Thorac Cardiovasc Surg. 1996; 111: 1267-1279Abstract Full Text Full Text PDF PubMed Scopus (117) Google Scholar, 2Murkin JM Martzke JS Buchan AM et al.A randomized study of the influence of perfusion technique and pH management strategy in 316 patients undergoing coronary artery bypass surgery: Part 2. Neurological and cognitive outcomes.J Thorac Cardiovasc Surg. 1995; 110: 349-362Abstract Full Text Full Text PDF PubMed Scopus (250) Google Scholar It is also likely that through the greater sensitivity introduced by standardizing preoperative cognitive performance between valve and CABG groups, Ebert et al were able to differentiate a significant difference in outcomes, even at these early postoperative intervals. This situation is in contradistinction to the findings of Andrew et al, in which the 6-month follow-up, rather than the 7-day assessment, showed decline in the valve group relative to the CABG group. The use of pH-stat management by Ebert et al for valve and CABG patients during cardiopulmonary bypass is of interest because an overall decrease in cognitive performance in patients subjected to pH-stat versus α-stat management during cardiopulmonary bypass has been convincingly shown in 2 separate clinical trials employing >400 CABG surgery patients.1Patel RL Turtle MR Chambers DJ et al.Alpha-stat acid-base regulation during cardiopulmonary bypass improves neuropsychologic outcome in patients undergoing coronary artery bypass grafting.J Thorac Cardiovasc Surg. 1996; 111: 1267-1279Abstract Full Text Full Text PDF PubMed Scopus (117) Google Scholar, 2Murkin JM Martzke JS Buchan AM et al.A randomized study of the influence of perfusion technique and pH management strategy in 316 patients undergoing coronary artery bypass surgery: Part 2. Neurological and cognitive outcomes.J Thorac Cardiovasc Surg. 1995; 110: 349-362Abstract Full Text Full Text PDF PubMed Scopus (250) Google Scholar This higher rate of injury has been shown to be due to the significantly higher numbers of cerebral emboli associated with pH-stat management, a consequence of the attendant cerebral hyperperfusion.1Patel RL Turtle MR Chambers DJ et al.Alpha-stat acid-base regulation during cardiopulmonary bypass improves neuropsychologic outcome in patients undergoing coronary artery bypass grafting.J Thorac Cardiovasc Surg. 1996; 111: 1267-1279Abstract Full Text Full Text PDF PubMed Scopus (117) Google Scholar Whether the use of pH-stat management had a relatively greater effect on the valve group in this study, given the exponentially higher rates of cerebral embolization associated with heart valve procedures, is unclear. Ebert et al postulate an anatomic basis for the observed cognitive dysfunction (ie, memory performance) imputing selective hippocampal vulnerability to hypoxia as a possible mechanism. In considering this possibility, Ebert et al are well in line with the thinking of clinicians who consider the evidence for cerebral microemboli, particularly as associated with use of cardiotomy suction,6Brooker JF Brown WR Moody DM et al.Cardiotomy suction: A major source of brain lipid emboli during cardiopulmonary bypass.Ann Thorac Surg. 1998; 65: 1651-1655Abstract Full Text Full Text PDF PubMed Scopus (221) Google Scholar to be consistent with this observation. It appears most likely that microemboli-induced regional cerebral ischemia, with secondary regional hypoxia, is at root cause, rather than systemic hypoxia per se. Given that patients undergoing heart valve surgery suffer exponentially more emboli than do patients undergoing CABG surgery, predominantly of a microgaseous nature, this strongly imputes a significant negative outcome to microgaseous air embolization. In summary, the efforts of all these groups are to be applauded. The time, effort, and degree of commitment necessary to undertake any such clinical studies in this complex area are huge. Andrew et al continue to establish and employ important new methodologic standards for this line of investigation. Their study and that of Ebert et al are further evidence for the putative role of microgaseous air in the genesis of some forms of postoperative brain injury. Silbert et al are to be encouraged in their efforts. Early postoperative cognitive performance is a potentially important and sensitive discriminator to evaluate new techniques and technologies. Given the experiences of Silbert et al, perhaps one meaningful endpoint in the early postoperative period is actual numbers of patients able to successfully complete such postoperative cognitive testing per se, as analyzed for anesthetic technique or other types of interventions. Attention to methodologic details, and incorporation of control group-derived reliable change indices can only enhance further understanding of these issues.