| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2004;78:1562-1563
© 2004 The Society of Thoracic Surgeons
University Medical Center Utrecht, Department of Anesthesiology Heidelberglaan 100, Utrecht, The Netherlands
ddijk{at}azu.nl
Several studies have reported an association between aortic arteriosclerosis and the number of cerebral microemboli, detected by transcranial Doppler. The authors' hypothesis, that aortic disease not only predicts stroke but also neurocognitive dysfunction (NCD), is therefore very reasonable. They carried out a study in a substantial sample of 162 patients and measured cognitive performance at relevant time points. Yet, this study failed to establish such an association.
Retrospective studies may be carried out without a protocol and a clear analytical plan, designed before exploration of the data. Therefore, a large number of analyses are often performed. When one of these confirms a hypothesis, it is tempting to present this particular analysis, while ignoring others. In contrast, if none of the many analyses confirms the hypothesis, this fact may be presented as the ultimate proof that there is really no association. In the present study, no association was found between (on one hand) atheroma in the ascending aorta, or the aortic arch, or the descending aorta (defined as a continuous variable, or as a binary variable), and (on the other hand) the cognitive change score, or a decline in a cognitive domain.
How good is this evidence? Is it still conceivable that aortic arteriosclerosis leads to NCD, despite the fact that all analyses in the present study failed to demonstrate such an association? Is the absence of evidence, evidence of absence? Perhaps there are other explanations. One is that the study may still have been underpowered to detect a relatively small correlation. The authors did excellent work by performing a sensitivity analysis to determine the largest effect size that their analyses may have missed due to insufficient power. Only a maximum of 6.25% of the variability in cognitive performance could be caused by atheroma burden, without being detected in this patient sample, suggesting that this study has sufficient power. But what if a large part of the remaining variability is caused by noise? It is striking that healthy subjects undergoing repeated cognitive testing have substantial natural fluctuations in their performance. The commonly used definitions of cognitive decline therefore generate high numbers of false positive outcomes and the incidence of NCD after bypass surgery is probably significantly overestimated [1]. Measuring a correlation with a rare event may require a sample that is much larger than 162 patients.
Even if there remains uncertainty about the role of aortic atheroma burden and the development of NCD, there is no doubt that aortic atheroma increases the risk of stroke. Manipulation of a patient's ascending aorta remains a dangerous maneuver.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
