Letter by Rubens et al Regarding Article, “Continuous-Flow Cell Saver Reduces Cognitive Decline in Elderly Patients After Coronary Bypass Surgery”
To the Editor:
We read with interest the publication of Djaiani et al addressing the effect of cardiotomy blood processing with a continuous-flow cell saver on the incidence of postoperative cognitive deficits (POCD) in the elderly undergoing coronary artery bypass grafting.1 The authors were able to demonstrate a significant decrease in the incidence of POCD with this intervention.
We have recently published our larger double-blind randomized clinical trial (266 patients) examining this same question.2 We committed a priori to a commonly used definition of what constitutes POCD, and we were not able to demonstrate a benefit either early postoperatively or after 3 months. On the other hand, we documented an increased risk of bleeding and transfusion associated with the intervention. Our confidence in this finding was strengthened by the blinding of all the participants as well as by the high rate of compliance to a strict transfusion protocol in both groups.
Because the widespread adoption of this practice—in the hope of decreasing the incidence of POCD—will result in increased blood loss and transfusion, it is important to critically examine the methodology of these 2 trials with opposing results. Readers and reviewers of Circulation may not be familiar with the problems of analyzing and interpreting psychometric data in a repeated-measures design. The most commonly used approach defines a subject as having suffered a deficit if scores decrease at least 20% in at least 20% of the tests administered.3 Another accepted approach involves combining tests into 3 or 4 independent cognitive domains using z scores4 or principal components analysis.5 A patient is then considered to have suffered a deficit if at least a 1 SD decrease is present in at least 1 domain. The method of Rasmussen6 referenced by Djaiani et al is well conceived but cannot be applied to their data because it requires that researchers use data from nonsurgical controls tested at the same interval as the subjects in order to correct for learning effects associated with repeated testing. Also, this method requires a decline of at least 2 SD in the combined score, not 1 SD as described by Djaiani et al.
Because of the unique nature of the analysis presented, we find it difficult to compare these results with current studies in the literature. We would ask that a standard analysis be reported (ie, the 20/20 rule, the method of Van Dijk,7 quoted but not presented: a decrease of at least 1 SD in the test z score in at least 20% of the tests), or that the domains constructed by the method of Townes or Newman be used.
Having chosen a priori an accepted definition of POCD, we found that cardiotomy blood processing does not protect the brain, but it does make our patients bleed and it exposes them to excessive blood products. Combining our data with those of Djaiani and using a standard analysis on the combined data set may be helpful in more accurately answering this important question relevant to the care of patients undergoing cardiac surgery.
Djaiani G, Fedorko L, Borger MA, Green R, Carroll J, Marcon M, Karski J. Continuous-flow cell saver reduces cognitive decline in elderly patients after coronary bypass surgery. Circulation. 2007; 116: 1888–1895.
Rubens FD, Boodhwani M, Mesana T, Wozny D, Wells G, Nathan HJ. The cardiotomy trial: a randomized, double-blind study to assess the effect of processing of shed blood during cardiopulmonary bypass on transfusion and neurocognitive function. Circulation. 2007; 116 (suppl): I-89–I-97.
Van Dijk D, Jansen EW, Hijman R, Nierich AP, Diephuis JC, Moons KG, Lahpor JR, Borst C, Keizer AM, Nathoe HM, Grobbee DE, De Jaegere PP, Kalkman CJ. Cognitive outcome after off-pump and on-pump coronary artery bypass graft surgery: a randomized trial. JAMA. 2002; 287: 1405–1412.