Free
Correspondence  |   December 2005
Graphical Display of Data Could Reveal Errors in Statistical Tests
Author Affiliations & Notes
  • Michael G. Irwin, M.B.Ch.B., M.D., F.R.C.A., F.H.K.C.A., F.H.K.A.M.
    AMPERSANDNUMBERSIGNx0002A;
  • AMPERSANDNUMBERSIGNx0002A;University of Hong Kong, Hong Kong.
Article Information
Correspondence
Correspondence   |   December 2005
Graphical Display of Data Could Reveal Errors in Statistical Tests
Anesthesiology 12 2005, Vol.103, 1315. doi:0000542-200512000-00032
Anesthesiology 12 2005, Vol.103, 1315. doi:0000542-200512000-00032
In Reply:AMPERSANDNUMBERSIGNx02014;
We thank Dr. Guignard for his comments and interest in our article. We have checked figure 1 as presented in the text and found that there was an error in the graphical presentation of this figure. There were higher behavioral pain scores in the first 10 min in patients who had received remifentanil but not at 15 min and, as Dr. Guignard correctly points out, this has not been made clear in the text. For completeness, the tabular data are presented below (), with our statistical analysis using the continuity adjusted chi-square test to analyze the behavioral pain score (statistical software: SAS System for Windows Release 8.02; SAS Institute Inc., Cary, NC).
Table 1. Numbers of Patients Classified by Behavioral Pain Scores 
Image not available
Table 1. Numbers of Patients Classified by Behavioral Pain Scores 
×
There was no difference by 15 min and no difference in visual analog pain scale scores. We believe that these differences in the first 10 min relate to pharmacokinetic differences between remifentanil, nitrous oxide, and the titration of drugs at the end of the case because the scores rapidly equilibrated and they are of little significance compared with the main outcome measures of this study. In retrospect, a narrow scoring system such as this may also have limitations in discriminating differences.
There was no difference in the total morphine consumption or the time to the first dose of morphine during the stay in the recovery room. The two groups had similar total morphine consumption in the first 24 h and visual analog pain scale scores at rest and movement. The reported incidence of postoperative nausea and vomiting was 10% in both groups. There was no difference in the sedation scores.
Our main objective was to determine whether the substitution of remifentanil for nitrous oxide, an increasingly common clinical practice, results in acute opioid tolerance. To more tightly control this study, we had to substitute remifentanil for nitrous oxide, as far as possible, while otherwise maintaining a normal standard of care. At our institution, that involves fentanyl coinduction and morphine before skin incision. This is why this occurs in both groups.
We apologize for this oversight in the behavioral pain score data presentation but are confident that this does not detract from the principal conclusions of the study.
AMPERSANDNUMBERSIGNx0002A;University of Hong Kong, Hong Kong.
Reference
Reference
Lee L, Irwin M, Lui SK: Intraoper1ative remifentanil infusion does not increase postoperative opioid consumption compared with 70% nitrous oxide. Anesthesiology 2005; 102:398AMPERSANDNUMBERSIGNx02013;402Lee, L Irwin, M Lui, SK
Table 1. Numbers of Patients Classified by Behavioral Pain Scores 
Image not available
Table 1. Numbers of Patients Classified by Behavioral Pain Scores 
×