To the Editor:—
Firth and Head1are to be congratulated for providing a comprehensive review of the pathophysiology of sickle cell disease and erudite discussion of its implications for anesthesiologists. We also appreciate their attempts to apply evidence-based knowledge to our understanding of the perioperative care of these patients. Although this will undoubtedly become a valuable resource for anesthesiologists, we are compelled to provide some corrections in the text and tables as well as provide alternative interpretations of some of the evidence.
The authors refer to a randomized no-transfusion group in their discussion of Vichinsky et al. (reference 98, page 776).2In fact, there was not a randomized no-transfusion group. The authors seemed to have missed the fact that this report and several more by the same authors were reports of subpopulations of the work by this group. In this and subsequent articles from the Preoperative Transfusion in Sickle Cell Disease Study Group as initially reported by Vichinsky et al. ,3the groups were the same (described below). Furthermore, the second to last sentence in the third paragraph on page 776 should read, “This [acute chest syndrome] occurred in 21% of cases in both the aggressive transfusion and the nonrandomized nontransfusion group, 8% in the conservative transfusion group, and 3% in the nonrandomized transfusion group,” as described in table 4.
In table 4 on page 775, regarding Haberkern et al., “1995,”4the numbers and percents for “cholecystectomies, complications” for the four groups are in fact the numbers of patients in the groups and the percents of sickle cell events, not the numbers and percents of complications. (This study was actually published in 1997.) The percentages of total complications and acute chest syndrome in the four groups (as listed) are in fact as follows: group 1, randomized aggressive transfusion: 36%/9%; group 2, randomized simple transfusion: 39%/11%; group 3, nonrandomized nontransfusion: 43%/19%; and group 4, nonrandomized transfusion: 41%/7%. These groups are the same in all of the studies reported by the Preoperative Transfusion in Sickle Cell Disease Study Group. These corrected data underscore concerns regarding the risk of perioperative complications in the nontransfusion group.
In the discussion of Griffin and Buchanan,5the authors concluded that “any potential benefit from transfusion would therefore be low and risks of transfusion were not justified for minor procedures.” However, the actual conclusion from this report stated that “operative transfusions might be avoided in children with sickle cell disease who undergo most minor surgical procedures.” The overall complication rate was 26%, thoracotomy/laparotomy 50%, tonsillectomy and adenoidectomy 56%, others 5%. This report neither provided evidence to withhold transfusion in any group nor lobbied against transfusions.
In their table 5 on page 777 (Guidelines for the use of Perioperative Prophylactic Erythrocyte Transfusion), the foundation for this table is not clear and certainly not evidence based. It suggests guidelines for perioperative transfusion that are misleading given the absence of prospective, randomized data to support a nontransfusion approach.
We acknowledge the lack of a proven causal relation between hypoxia, dehydration, and hypothermia and sickling events in the perioperative clinical setting (page 782). However, in the context of sickle cell disease, we are compelled to prove that no such relation exists before abandoning practices that have been associated with decreased perioperative morbidity and mortality in these patients.
A conservative approach to children with sickle cell disease in the perioperative period has been and continues to be adequate hydration and correction of anemia. To propose a therapeutic nihilistic approach to the treatment of these patients in the absence of substantive evidence is dangerous. Despite the best care today, the perioperative mortality rate in patients with sickle cell disease of 1 in 100 is severalfold greater than that in nonsickle adults, approximately 1 in 300,000, and in nonsickle children, 1 in 50,000–80,000. Before our current practice patterns for these children are changed, prospective randomized studies that examine anesthetic practices in this and other diseases should be conducted.
* Nemours Children's Clinic, Jacksonville, Florida. firstname.lastname@example.org