Lenoir Brigitte, Merckx Paul, Paugam-Burtz Catherine, Dauzac Cyril, Agostini Marie-Madeleine, Guigui Pierre, Mantz Jean
Department of Anesthesia and Critical Care, Beaujon University Hospital, Clichy, France.
Anesthesiology. 2009 May;110(5):1050-60. doi: 10.1097/ALN.0b013e31819df9e0.
The aim of this study was to generate a score based on preoperative characteristics and predictive of the individual probability of allogeneic erythrocyte transfusion in patients undergoing elective thoracolumbar spine surgery.
Two hundred thirty consecutive patients were retrospectively included over a 15-month period (derivation set). Preoperative independent predictors of erythrocyte transfusion from the day of surgery until postoperative day 5 were determined by multivariable analysis, from which a model of individual probability of transfusion was derived and prospectively validated in 125 additional patients (validation set).
Four preoperative independent predictors were associated with transfusion: age older than 50 yr (adjusted odds ratio = 4.9 [2-13.5]), preoperative hemoglobin level less than 12 g/dl (adjusted odds ratio = 6.9 [3.1-17.2]), fusion of more than two levels (adjusted odds ratio = 6.7 [3.1-15.2]), and transpedicular osteotomy (adjusted odds ratio = 19.9 [5.6-98.2]). A 0-4 score (0 = no risk, 4 = maximum risk) predictive of allogeneic transfusion was derived by weighting estimate parameters for each variable in a multivariable logistic regression model. Discriminating capacity of the score was 0.86 [0.81-0.92] in the receiver operating characteristics in the derivation sample and 0.83 [0.75-0.91] in the validation sample. The observed transfusion rates in the validation set and the individual probabilities of erythrocyte transfusion from the score were well correlated (y = 0.98x + 0.04; P < 0.0001), and the observed differences were not statistically different (goodness-of-fit chi-square, P = 0.125). The score was also correlated with the number of erythrocyte units transfused (Spearman rho = 0.61; P < 0.0001).
The Predictive Model of Transfusion in Spine Surgery may be useful in clinical practice to identify patients undergoing spine surgery at risk of massive bleeding and encourage erythrocyte-saving strategies in these patients.
本研究的目的是基于术前特征生成一个评分，以预测择期胸腰椎手术患者接受异体红细胞输血的个体概率。
在15个月的时间里回顾性纳入了230例连续患者（推导集）。通过多变量分析确定从手术当天至术后第5天红细胞输血的术前独立预测因素，据此推导出血液输注个体概率模型，并在另外125例患者中进行前瞻性验证（验证集）。
四个术前独立预测因素与输血相关：年龄大于50岁（调整后的优势比 = 4.9 [2 - 13.5]）、术前血红蛋白水平低于12 g/dl（调整后的优势比 = 6.9 [3.1 - 17.2]）、融合节段超过两个（调整后的优势比 = 6.7 [3.1 - 15.2]）以及经椎弓根截骨术（调整后的优势比 = 19.9 [5.6 - 98.2]）。通过对多变量逻辑回归模型中每个变量的估计参数进行加权，得出了一个预测异体输血的0 - 4分（0 = 无风险，4 = 最大风险）评分。该评分在推导样本的受试者工作特征曲线中的鉴别能力为0.86 [0.81 - 0.92]，在验证样本中为0.83 [0.75 - 0.91]。验证集中观察到的输血率与根据评分得出的红细胞输血个体概率高度相关（y = 0.98x + 0.04；P < 0.0001），且观察到的差异无统计学意义（拟合优度卡方检验，P = 0.125）。该评分还与输注的红细胞单位数量相关（Spearman相关系数rho = 0.61；P < 0.0001）。
脊柱手术输血预测模型在临床实践中可能有助于识别有大出血风险的脊柱手术患者，并鼓励对这些患者采取节约红细胞的策略。
