Durand Marie-Christine, Porcher Raphaël, Orlikowski David, Aboab Jérôme, Devaux Christian, Clair Bernard, Annane Djillali, Gaillard Jean-Louis, Lofaso Frédéric, Raphael Jean-Claude, Sharshar Tarek
Functional Testing Department, Raymond Poincaré Teaching Hospital, Garches, France.
Lancet Neurol. 2006 Dec;5(12):1021-8. doi: 10.1016/S1474-4422(06)70603-2.
Respiratory failure is the most serious short-term complication of Guillain-Barré syndrome and can require invasive mechanical ventilation in 20-30% of patients. We sought to identify clinical and electrophysiological predictors of respiratory failure in the disease.
We prospectively assessed electrophysiological data and clinical factors, including identified predictors of delay between disease onset and admission, inability to lift head, and vital capacity, in patients admitted with Guillain-Barré syndrome. We related these factors to subsequent need for ventilatory support. Neurophysiological findings were classified as demyelinating, axonal, equivocal, unexcitable, or normal. Predictive values of clinical and electrophysiological data were tested using classification trees to build up a predictive model. This model was initially built up in a two-third (fitting set) then validated in a one-third (validation set) of the total sample. The fitting and validation sets were randomly selected. We also assessed the predictive value of this model for disability at 6 months.
From 1998, to 2006, 154 patients with Guillain-Barré syndrome were included in the study and 34 (22%) were subsequently ventilated. Demyelinating Guillain-Barré syndrome was more common in patients who went on to be ventilated than in those who were not (85%vs 51%, p=0.0003). Vital capacity and the proximal/distal compound muscular amplitude potential (p/dCMAP) ratio of the common peroneal nerve were retained in the tree model, with a probability of needing ventilation of less than 2.5% in patients with a ratio of greater than 55.6% and a vital capacity more than 81% of predicted. A p/dCMAP ratio of the peroneal nerve less than 55.6% and age older than 40 years were retained as independent predictors of disability at 6 months.
Neurophysiological testing is helpful for assessing risk of respiratory failure, which is highest in patients with evidence of demyelination and very low in those without both 55.6% conduction block of the common peroneal nerve and a 20% reduction in vital capacity.
呼吸衰竭是吉兰-巴雷综合征最严重的短期并发症，20% - 30%的患者可能需要有创机械通气。我们试图确定该疾病中呼吸衰竭的临床和电生理预测因素。
我们前瞻性地评估了吉兰-巴雷综合征患者的电生理数据和临床因素，包括已确定的疾病发作与入院之间延迟的预测因素、无法抬头以及肺活量。我们将这些因素与随后对通气支持的需求相关联。神经生理学发现被分类为脱髓鞘型、轴索性、不明确型、无兴奋性型或正常型。使用分类树测试临床和电生理数据的预测价值，以建立预测模型。该模型最初在总样本的三分之二（拟合集）中建立，然后在三分之一（验证集）中进行验证。拟合集和验证集是随机选择的。我们还评估了该模型对6个月时残疾情况的预测价值。
从1998年到2006年，154例吉兰-巴雷综合征患者纳入研究，其中34例（22%）随后接受了通气治疗。脱髓鞘型吉兰-巴雷综合征在接受通气治疗的患者中比未接受通气治疗的患者更常见（85%对51%，p = 0.0003）。树模型中保留了肺活量和腓总神经的近端/远端复合肌肉动作电位（p/dCMAP）比值，当比值大于55.6%且肺活量超过预测值的81%时，患者需要通气的概率小于2.5%。腓总神经的p/dCMAP比值小于55.6%和年龄大于40岁被保留为6个月时残疾的独立预测因素。
神经生理学检测有助于评估呼吸衰竭风险，在有脱髓鞘证据的患者中风险最高，而在既没有腓总神经55.6%传导阻滞也没有肺活量降低20%的患者中风险非常低。
