Skip Navigation LinksHome > June 2005 - Volume 100 - Issue 6 > Failing to Reject the Null Hypothesis Does Not Mean that the...
Anesthesia & Analgesia:
doi: 10.1213/01.ANE.0000156679.52332.42
Letters to the Editor: Letters & Announcements

Failing to Reject the Null Hypothesis Does Not Mean that the Null Hypothesis Is True

Schwann, Nanette M. MD; Horrow, Jay MD, MS

Free Access
Article Outline
Collapse Box

Author Information

Allentown Anesthesia Associates, Vice Chair for Research, Department of Anesthesiology, Lehigh Valley Hospital and Health Network, Allentown, PA, Nan.schwann@lvh.com (Schwann)

Department of Anesthesiology, Drexel University College of Medicine, Philadelphia, PA (Horrow)

Back to Top | Article Outline

In Response:

The authors agree with Dr. McIlroy’s assessment of the limitations of these data and clearly state in the article “. . .our data suggest that avoidance of CPB does not reduce renal morbidity. . .” Inexplicably, Dr. McIlroy’s quotation missed the important qualifier “suggest.” Moreover, the data are observational and not experimental, i.e., they lack randomization, and thus permit conclusions only with regard to association rather than causation. A trial enrolling even the 5700 patients Dr. McIlroy desires to detect his arbitrarily selected 25% reduction in increased serum creatinine (even if it were feasible) would still allow no inference regarding the impact of cardiopulmonary bypass on renal dysfunction. Alternatively, an analysis using a surrogate marker like glomerular filtration rate may detect smaller differences between off-pump coronary artery bypass and coronary artery bypass grafting using fewer patients, but, as Dr. McIlroy points out, these markers are less “clinically relevant” (less clearly associated with negative outcome).

Because the incidence of serious end-organ complications after coronary revascularization (excluding Type II neurological dysfunction) remains less than 5%, Type II error remains a serious concern in all investigations enrolling less than several thousand patients. Ultimately, the superiority of either on-pump or off-pump revascularization will require a large randomized trial, ideally sponsored by the National Institutes of Health. Our published report (and others including a large meta-analysis soon to be published [personal communication]) calls for such an adequately powered trial. Without such a trial, the superiority of one technique over another will remain conjecture.

Finally, McIlroy ponders “the ethics behind the use of funding for a trial in which a negative result will be meaningless.” What funding source does he wish to protect from this putative unethical practice? The authors have no relevant financial or consulting relationships to disclose and conducted this analysis as an investigator-initiated unfunded trial in their “spare” time. Would he have raised the same issue had the results shown a difference?

Nanette M. Schwann, MD

Allentown Anesthesia Associates

Vice Chair for Research

Department of Anesthesiology

Lehigh Valley Hospital and Health Network

Allentown, PA

Nan.schwann@lvh.com

Jay Horrow, MD, MS

Department of Anesthesiology

Drexel University College of Medicine

Philadelphia, PA

© 2005 International Anesthesia Research Society

Login

Become a Society Member