| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
From the *Department of Anesthesiology and Critical Care Medicine; and Department of Pediatrics, Johns Hopkins School of Medicine, Baltimore, Maryland.
Address correspondence and reprint requests to Myron Yaster, MD, The Johns Hopkins Hospital, 935 Blalock, 600 N. Wolfe Street, Baltimore, MD 21287. e-mail to myaster{at}jhmi.edu
The decision of whether to anesthetize a child with an acute upper respiratory tract infection (URI) is among the most common problems faced by the anesthesiologist who treats children. Not long ago, conventional wisdom dictated that these children were at an increased risk of developing intra- and/or postoperative complications and were routinely cancelled for elective surgery (and anesthesia). But like many other "truths" that have not stood the test of time and evidence (e.g., transfusing for hemoglobin levels of <10 g/dL or NPO for liquids after midnight), most anesthesiologists practicing today do not consider an uncomplicated URI (without evidence of lower airway disease) to be a contraindication to surgery (1,2). The exception may be the child with congenital heart disease presenting for cardiac surgery (3). In this situation, postponement of cardiac surgery for approximately six weeks is advisable to allow airway inflammation to subside. Unfortunately, this is not always possible, because the timing of congenital heart surgery is often dictated by the physiology of the particular presenting lesion and is often not elective. Even with this timing constraint, modern perioperative care has allowed pediatric cardiac patients with URI to experience the same survival rates and total hospital lengths of stay as children without URI (3).
Despite these successes in managing potentially life-threatening complications in the operating room (OR), a simple, readily available preventive therapy would enhance perioperative safety and make the anesthesiologist's job easier. Only four randomized controlled trials have investigated the role of therapy for the prevention or treatment of the anesthetized child with URI. Collectively, they have found that a laryngeal mask airway is safer than an endotracheal tube for children with URI (4), that sevoflurane and halothane have equivalent complication rates (5) and that neither inhaled albuterol nor inhaled ipratropium offer protection against adverse intraoperative events in children with URI (6).
In this issue of Anesthesia & Analgesia, Tait et al. (7) add to this literature by investigating whether glycopyrrolate is effective for preventing intraoperative complications, such as bronchospasm, laryngospasm, or hypoxia, among children with URI. The authors found that glycopyrrolate did not decrease respiratory complications in children with URI. While apparently not affecting the incidence of potentially life-threatening intraoperative complications, glycopyrrolate did lead to an increased incidence of self-limited postoperative complications, such as wheezing, dehydration, dizziness, facial flushing, and hyperactivity, which were offset by the reduced incidence of postoperative nausea and vomiting (PONV) (7).
The difficulties of research design and execution with this and all single institution studies of complex, multifactorial phenomena in the OR are formidable. First, there is the large barrier of assembling a sufficient sample size. In the Tait et al. (7) study the sample size was only adequate to detect a >50% reduction in complication rate. Very few drugs have a 50% success rate for treating complex, multifactorial life-threatening problems in the operating room or intensive care unit. A 20% treatment effect would be considered a more reasonable criterion for successful treatment of complex, multifactorial responses such as intraoperative cough, hypoxia, breath-holding, laryngospasm, or bronchospasm. However, this would require a sample size of approximately 320 patients per group using the same assumptions as the authors did, namely of a 51% incidence of perioperative respiratory adverse events in children with URIs. If the baseline incidence of complications with URI were only 25%, the sample size requirement increases to more than 1000 patients (1094) per group. No single children's hospital in the United States is likely to generate sample sizes of this magnitude for a randomized controlled trial. Other design difficulties include the inability to blind the anesthesiologist to the treatment group and removal of otherwise eligible patients because of parent, surgeon, or anesthesiologist's preference. As the authors acknowledge, the conclusion that glycopyrrolate has no effect on adverse respiratory event in children with URI must be tempered by these limitations of the study design.
This study underlines the need for a multi-institutional collaborative effort and external funding. The days of large-scale, randomized, controlled trials using funds from only a single department of anesthesiology are probably over. There may have been many multi-institutional research efforts in pediatric anesthesiology over the years, but they have been largely ad hoc, and either unfunded or funded by pharmaceutical or device companies. Clinical research conducted in partnership with industry has typically focused on the pharmacology of new anesthetic drugs or the efficacy of new devices. In addition, pediatric anesthesia research has enhanced the management of postoperative pain and PONV. All of these areas will remain an important component of the research portfolio in pediatric anesthesia. However, unless we adopt some new strategies, the type of study attempted by Tait et al. in this issue; namely, prevention and management of complex, potentially life threatening events in the OR, will fade.
As illustrated by the Tait et al.'s study, the greatest hurdle is achieving adequate sample size. In this regard, perhaps pediatric anesthesiology is a victim of its own success, in that survival after anesthesia is no longer a real issue. The most recent large-scale study was published in 2002 and showed an anesthesia-related mortality rate of 0.36 per 10,000 (8,9). However, even these data come from an era in which halothane was still in use for infants and accounted for 37% of the mortality in the study. Now that halothane is no longer in clinical use in the United States, one might expect that the anesthesia-attributable mortality is even lower. Hence, mortality is not a viable end-point for prospective studies in pediatric anesthesia.
The only solution to the sample size dilemma is to get beyond the confines of the single institution and develop research consortia in pediatric anesthesia. Research consortia coalesce because of the perception of a critical problem that requires collaboration to answer. To our knowledge there is only one example of a research consortium devoted solely to a pediatric anesthesia problem. The Pediatric Sedation Research Consortium (10) has logged more than 30,000 sedation encounters from 26 institutions and found critical complications from pediatric sedation to be very rare.
If we agree that multicentered trials are needed with external funding, then this raises an even broader question for the entire subspecialty. What are the "big picture" questions that face the subspecialty, and how and where can studies be performed to answer these questions? The national focus on patient safety and the desire for high-quality pediatric anesthesia research beyond an industry-sponsored research agenda mean that pediatric anesthesia will have to embrace new research questions and research designs. For example, while the short-term outcomes after anesthesia or life-threatening perioperative event appear excellent, very little is known about long-term outcomes. Using the Veterans Administration and the multiple studies of adults with coronary artery disease undergoing noncardiac surgery as an example, it is now very clear that the use (or failure to provide) ß- blockers or clonidine in the preoperative period may affect mortality up to 2 yr after surgery and anesthesia (11,12). These findings are nothing short of earth shaking. In pediatric anesthesia, it is unclear what the long-term neurodevelopmental effects are of anesthetizing newborn infants with general anesthetics, or if glucose control is important, or what, if any, are the effects of short- and long-term pain and its management on the central nervous system (13,14). Another example is the impact of massive transfusion requirements on the failure to wean from mechanical ventilation and subsequent chronic ventilator dependency after spinal fusion. If pilot studies confirm the possibility of disability induced by anesthetic or intraoperative events, then research consortia might attract external funding to examine complex intraoperative phenomena.
The standards of evidence-based medicine have considered the randomized controlled trial as the "gold standard." Anyone who has attempted to apply this research design to complex intraoperative phenomena can attest to the challenges. The factorial design offers a possible alternative, since it allows analysis of two or more independent variables (or interventions) in a single study in an efficient manner that permits an understanding of the interactions among the independent variables (15). For instance, Apfel et al. (16) found that sixcombinations of antiemetics and anesthetics were equivalent in preventing PONV. With six treatment combinations, there were a possible 64 (26) treatment combinations (blocks), assuming all treatment combinations were equally effective. Not all participating centers had to study all 64 treatment combinations. Rather, an individual center was assigned a maximum of four treatment combinations. While such an assignment is ideally undertaken randomly (as it was in the Apfel et al. study), a targeted assignment of treatment combinations based on local preferences could begin to address one of the major impediments to large-scale collaborative research in pediatric anesthesia.
Complex intraoperative events should remain an area of important inquiry for the pediatric anesthesiologist despite the real successes in improving patient safety, but some adaptations are needed. Research consortia will allow larger sample size. An examination of longer-term outcomes will provide new questions around which these consortia might coalesce. And, finally, new research design methods can help overcome some of the methodologic barriers to understanding complex intraoperative phenomena.
 |
Footnotes |
|---|
|
REFERENCES |
|---|
 Top REFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
