| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 2003;126:1660-1661
© 2003 The American Association for Thoracic Surgery
Letters to the editor |
a Department of Surgery, University of Illinois at Chicago, 840 Southwood St (CSB suite 518-E), Chicago, IL 60612, USA
To the Editor:
The article by Grunkemeier and colleagues1 published in the February 2003 issue of the Journal is clearly an important study demonstrating the role of logistic regression analysis performed on a series of statistical summaries from different institutions. The authors' approach toward such summary of studies may not, however, have offered the solution to the confusion present when trying to answer for differences in complication rates between prosthetic valves. They set out to offer explanations for the heterogeneity among the study series analyzed. However, they did not address any clinical context in their explanation for such diverse outcomes with regard to valve thrombosis, bleeding, and thromboembolism.
A number of questions need to be addressed before reaching such a conclusion from summary of a series. Three important aspects that were not mentioned include the etiology of the valve disease, case mix, and comorbidities. All three factors should be addressed in the context of which valve is superior (St Jude vs CarboMedics) in either mitral or aortic position to that unique patient's heart valve biology. No mention was made of operative variables, including in-hospital reoperative cases and estimated blood loss.
To replace a heart valve with a certain type of prosthesis is a gray zone in the field of surgery, with as yet little evidence-based medicine for a valid argument. Differential understanding, surgical expertise, and institutional experience still remain the main determining factors for definitive decision, except in a clinical trial. A potential drawback of analysis of summary statistics in this case is that it nearly forces the investigator to ask and attempt to answer the wrong question ('Is it better to replace with St Jude or a CarboMedics valve?') rather than to ask who can undergo replacement and, among those, who will benefit from replacement. This requires analysis of individual patient data considered more powerful under all circumstances. It may be more difficult to retrieve such data because it requires the cooperation of several institutions and individuals and determination of common definitions or adjudication thereof. In a similar context, Lim and coworkers2 in a randomized controlled trial showed that no difference existed between the two valve brands.
When recruiting studies that specifically address the question of mitral or aortic valve replacement, strict and valid inclusion criteria need to be adopted. The articles should include a patient population description, etiology of disease process, level of comorbidity, and surgical technique that are specific and well defined. In other words, the statistical summary should detail type of surgical procedure performed for a specific etiology. Other aspects that should be taken into account include whether the addition of coronary artery bypass or arrhythmia surgery or another valve replacement was performed. Moreover, a more valid and relevant postoperative follow-up will be achieved when including case series that address in detail preoperative assessment of mitral or aortic valve pathophysiology.
Before performing a logistic regression, four essential check points could have been taken into account. First is the etiology of mitral and aortic valve disease. Just as the pathophysiology differs among rheumatic, ischemic, degenerative, and infectious disease, so does the outcome. Ischemic etiology poses a particularly vexing problem. Ischemic normal mitral valve is highly variable. Therefore it is not surprising that results of replacement with what type of prosthesis are variable and controversial. For example, ischemic mitral valve disease is associated more often with diabetes. Diabetes is associated with a hypercoagulable state.3 So when replacing with either type of prosthesis, the outcome needs to be adjusted for the etiology, not forgetting the comorbidity. The second main point that needs to be addressed is how the subvalvular tensor apparatus was managed at operation. Left ventricular function is unquestionably related to this. At least in the context of randomized controlled trials, complete preservation of the apparatus results in better left ventricular performance,4 which also means a lower chance of thrombus and later embolism development.5 This generates another confounding factor when interpreting the data. Third, the underlying strategy for anticoagulation control should be addressed and controlled for in the inclusion criteria. Finally, the presence or absence of perioperative atrial fibrillation is acknowledged as important by both operating surgeons manufacturers.
It is possible that these factors were considered; however, for the sake of the reader, it is necessary to be include them in the discussion. Such consideration will reduce chance-related outcome and broaden the horizon of the reader when transferring a statistical critique to clinical reality.
References
Related Article
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
