Chart review: a blessing or a curse?

Neurological injury following cardiac surgery is a common complication that has a wide range of clinical presentations. Patients can present with subtle cognitive defects that can only be detected by carefully conducted neuropsychiatric testing. In contrast, some patients experience serious neurological injuries that result in death or disability. Many of the injuries can have profound social and economic impacts, regardless of how successful the cardiac surgery is. A timely diagnosis of central nervous system injury following cardiac surgery requires a comprehensive and concise neurological evaluation. Any individual change in performance from baseline following surgery is an important clue that should prompt a full evaluation. A patient with clear consciousness before surgery should have an uneventful recovery following uncomplicated cardiac surgery. Delayed emergence after cardiac surgery may be an indication of neurological damage. Thus, it is important to differentiate patients and surgical variables that can lead to delayed emergence. However, there is a lack of studies that focus upon the identification and evaluation of risk factors for delayed emergence after cardiac surgery. In this month’s issue of the Journal, Tsai et al report a retrospective study to examine the association between specific variables and emergence time after cardiac surgery. Although this is an important and noble task, there are numerous of potential confounders that make a rigorous analysis impossible. Some of these issues include patient characteristics, surgical technique, anesthetic doses, and intraoperative cardiac events. It is not an easy task to quantify all these factors or possible to include all imaginable confounders in a retrospective survey. The authors used a chart review to include as many potentially influential factors as they could in their study. Five risk factors of delayed emergence after cardiac surgery included old age, male gender, low body mass index, high preoperative blood urea nitrogen level, and prolonged cardiopulmonary bypass. Although the findings are interesting, there is an empirical quality to the study. The authors studied variables that have generally been associated with an increased risk of neurological injury. They were unable to determine what the predictive values of these factors were using their own data. Similar problems plague many retrospective studies. The precise identification of clinical correlations requires large databases or prospectively designed