Continuous Propofol Infusion in 142 Critically Ill Children
Barry P. Markovitz, MD, FAAPWashington University School of Medicine
St Louis Children’s Hospital
St Louis, MO 63110, USA
To the Editor.—
It is truly disappointing to be corresponding on this issue in the 21st century, but Cornfield et al1 continue to perpetuate the fallacy of developing a (false) sense of security by the lack of observed complications in a small study. In their review of 142 patients in their pediatric intensive care unit (PICU) receiving propofol by infusion, they found no complications attributable to the drug. Their conclusion that "propofol can be administered safely" ignores the concept of confidence intervals. In the eloquent article by Hanley and Lippman-Hand,2 the simple rule of thumb to calculate an upper 95% confidence interval for a zero complication rate is presented as 3/n. In Cornfield et al then, we can conclude with 95% certainty that the complication rate—which conceivably refers to the fatality rate with this drug—is no more than 2.1%? With thousands of PICU patients receiving this drug each year, this could potentially represent hundreds of unnecessary deaths. The medical literature is replete with examples of minor and major complications of drugs that were not detected until thousands of patients were studied.
The authors do not analyze this information in a vacuum and do include information from other sources to draw their conclusions. Nevertheless, there is no discussion of the statistical limitations of their sample size. Indeed, they portray this small, retrospective series as powerful enough to refute a recent Food and Drug Administration advisory regarding propofol for continuous sedation of PICU patients. This is most certainly not warranted.
REFERENCES
- Cornfield DN, Tegtmeyer K, Nelson MD, Milla CE, Sweeney M. Continuous propofol infusion in 142 critically ill children.
Pediatrics.2002; 110
:1177
–1181
[Abstract/Free Full Text] - Hanley JA, Lippman-Hand A. If nothing goes wrong, is everything all right?
JAMA.1983; 249
:1743
–1745
[Abstract/Free Full Text]
David N. Cornfield, MD*
Carlos E. Milla, MD*
Michael D. Nelson, MD*
Michael Sweeney, MD*,
* Division of Pediatric Pulmonology and Critical Care Medicine
Departments of Pediatrics and Anesthesia
University of Minnesota
Minneapolis, MN, USA
In Reply.—
The authors are neither surprised nor disappointed that even at the dawn of the 21st century intellectual purity must be balanced by rational pragmatism. Our conclusion that "propofol... can be safe and still effective, provided propofol dosing is carefully circumscribed" follows directly from the data presented.1 The report provides dosing guidelines for physicians choosing to use propofol in the course of caring for critically ill children in the pediatric intensive care unit (PICU). At no point in the article did the authors recommend the use of propofol.
Definitive demonstration of drug safety has been and always will be highly problematic, particularly in children. A trial designed to prove the absolute safety of propofol necessarily entails both the enrollment of an exceedingly large number of patients and the death of some patients enrolled in the trial. The relatively high background risk of serious complications in children admitted to PICUs (by definition PICU patients are critically ill), complicates the task of accurately determining the contribution of propofol to either death or serious complication in this patient population. Detecting with 90% confidence, a putative 50% increase in the incidence of death would require the enrollment of >7000 children treated with propofol and a similar number of children, matched with respect to both disease severity and age, not exposed to the drug.2 The sample size would preclude completion of the trial, as it represents a number that exceeds the entire patient population in residence in PICUs throughout North America at any given time. Perhaps more disturbing would be the need to demonstrate a fatality rate in children enrolled in the putative study. Enrolling children in a trial that anticipates the death of a discrete number of children, when a safer alternative is available, is not ethically acceptable.
Although in some areas of medicine a trial that includes 142 patients can justifiably be considered small, the present trial is not, in our estimation, such a case. The present report provides data on more children being administered propofol via continuous infusion in the PICU setting than the combination of all other reports that have appeared in peer-reviewed publications. Finally, the limitations of the study were carefully and fully discussed. Although the statistical limitations of the study were not explicitly addressed, the authors believe the readership is sufficiently familiar with standard statistical methods to allow for critical appraisal of the data and the conclusions.
REFERENCES
- Cornfield DN, Tegtmeyer K, Nelson MD, Milla CE, Sweeney M. Continuous propofol infusion in 142 critically ill children. Pediatrics.2002; 110 :1177 –1181
- Strom BL, ed. Pharmacoepidemiology. 2nd ed. Chichester, United Kingdom: Wiley; 1994
PEDIATRICS (ISSN 1098-4275). ©2003 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
