Reduction of transfusion rates in the surgical correction of sagittal synostosis

Object. As public concern about the risks of blood transfusions increased in the mid-1990s, avoidance of transfusions became a goal of surgery for sagittal synostosis. This study was performed to confirm a hypothesized reduction in transfusion rates in recent years and to identify factors associated with both the need for transfusion and low postoperative levels of hemoglobin. Methods. Sagittal synostosis operations performed in children between 1986 and 1999 were reviewed retrospectively. Patients underwent a minimum of vertex strip craniectomy and parietal craniectomies. There were 118 patients whose median age at surgery was 4.2 months. The primary end point for analysis was defined as either the receipt of a blood transfusion or a postoperative level of hemoglobin less than 70 g/L. Forty-two percent of patients (95% confidence interval [CI] 31-52%) treated before 1996 and 11% of patients (95% Cl 0-23%) treated from 1996 onward received blood. The reduction in the blood transfusion rate in later years was, in part, related to the acceptance of a lower postoperative hemoglobin level, often below 70 g/L. A univariate analysis showed that the only patient or surgical factors that correlated with reaching the primary end point in a statistically significant manner were the year of surgery and the extent of surgery. A logistic regression of the age and weight of the child, length of surgery time (from skin opening to skin closure), preoperative hemoglobin level, extent of surgery, and surgeon against the primary end point revealed that the best predictor of the need for a blood transfusion or the presence of a postoperative hemoglobin level lower than 70 g/L was the extent of surgery (beta = 1.4, standard error of the beta statistic = 0.44). Once the extent of surgery was accounted for in the model, no other covariates significantly improved the model. Techniques implemented to minimize blood loss since 1995 included the following: use of the Colorado needle for scalp incision, selection of the Midas Rex craniotome for cranial cuts, and application of microfibrillar collagen. Postoperative hemoglobin was allowed to decrease to 60 g/L if the child was stable hemodynamically, before blood was administered. There were no cardiovascular, wound healing, or infectious complications, and no surgeries were repeated for cosmetic reasons. Conclusions. Low blood transfusion rates were achieved using simple intraoperative techniques and by accepting a low level of postoperative hemoglobin.