Free
Correspondence  |   February 2009
Residual Confounding in Observational Studies
Author Affiliations & Notes
  • Jean-Luc Fellahi, M.D., Ph.D.
    *
  • *Centre Hospitalier Privé Saint-Martin, Caen, France.
Article Information
Correspondence
Correspondence   |   February 2009
Residual Confounding in Observational Studies
Anesthesiology 2 2009, Vol.110, 430. doi:10.1097/ALN.0b013e3181944aa3
Anesthesiology 2 2009, Vol.110, 430. doi:10.1097/ALN.0b013e3181944aa3
In Reply:—
We would like to thank Dr. Ho for his interest in our study.1 The propensity score used as a covariable in the second logistic regression model (covariance analysis) was not significantly associated with adverse cardiac outcome (P  < 0.15), as already mentioned in the footnote of table 4. Regardless of its P  value, we disagree with Dr. Ho that a significant propensity score would suggest residual confounding and the reference cited do not support this statement. Dr. Ho is correct when he notes that dobutamine was used predominantly in patients with a lower preoperative ejection fraction. This fact, however, does not necessarily mean that all these patients should have required inotropic support at the termination of cardiopulmonary bypass. The lack of cardiac monitoring to guide inotropic use is a major problem in daily practice, leading to possible inappropriate use of dobutamine for many patients undergoing cardiac surgery. Moreover, both EuroSCORE and left ventricular ejection fraction were relatively well balanced in table 3, which compared baseline characteristics of the subsample after matching groups on the propensity score.2 We believe that using multiple ways of propensity score adjustment with similar conclusions increases the internal validity of the present work. Of course, we agree with Dr. Ho that unknown confounders can limit the interpretation of any observational study, as we commented in the discussion section of the article, second paragraph.
At last, we would like to clarify an important remark pointed out by John Butterworth, M.D., in the Editorial View that accompanied our article.3 We have reported the results of a sensitivity analysis after exclusion of “those patients to whom nearly every clinician would administer a positive inotropic drug” with consistent results.
We have heard so many clinicians say regarding low doses of dobutamine: “If it does not help, it cannot hurt.” We believe it is high time to change clinical practices and elaborate guidelines for the perioperative use of inotropic agents in cardiac surgery. We also hope our study will stimulate research to confirm of refute our hypothesis-generating data, including replication in other populations. In this line, the results of our study now make the concept of a randomized trial ethically feasible.
*Centre Hospitalier Privé Saint-Martin, Caen, France.
References
Fellahi JL, Parienti JJ, Hanouz JL, Plaud B, Riou B, Ouattara A: Perioperative use of dobutamine in cardiac surgery and adverse cardiac outcome: Propensity-adjusted analyses. Anesthesiology 2008; 108:979–87Fellahi, JL Parienti, JJ Hanouz, JL Plaud, B Riou, B Ouattara, A
Austin PC: A critical appraisal of propensity-score matching in the medical literature between 1996 and 2003. Stat Med 2008; 27:2037–49Austin, PC
Butterworth J: Dobutamine: too dangerous for routine administration? Anesthesiology 2008; 108:973–4Butterworth, J