脑钠肽对脓毒性休克患者液体反应性的预测价值

Predictive value of B-type natriuretic peptide on fluid responsiveness in patients with septic shock

目的探讨血浆脑钠肽(BNP)能否作为一项反映心脏前负荷并预测液体反应性的指标,以指导脓毒性休克患者的容量复苏。方法采用前瞻性随机对照临床研究方法,纳入2011年1月至2015年12月海南省农垦总医院重症医学科78例脓毒性休克患者,纳入研究后立即给予容量复苏,容量复苏的方法为6%羟乙基淀粉300~500 m L/30 min静脉滴注,胸腔内血容量指数(ITBVI)>850 m L/m2为复苏终点,其他干预措施相同。采用脉搏指示连续心排血量监测仪(Pi CCO)监测患者容量复苏前后的血流动力学参数。以容量复苏后心排血量(CO)增加值(△CO)≥15%定义为容量反应阳性(扩容有效组),△CO<15%定义为容量反应阴性(扩容无效组)。评价两组患者容量复苏前后血流动力学参数、BNP及其变化值的差异,分析其与△CO的相关性。绘制受试者工作特征曲线(ROC),评价血流动力学参数、BNP预测容量反应性的价值。结果容量复苏前,扩容有效组和无效组间心率(HR)、平均动脉压(MAP)、中心静脉压(CVP)、心排血量(CO)、每搏输出量指数(SVI)、ITBVI、血管外肺水指数(EVLWI)比较差异均无统计学意义(P>0.05),但扩容有效组每搏变异度(SVV)明显高于扩容无效组[(13.05±4.75)%vs(9.71±4.38)%,P=0.034],BNP、全心舒张末期容积指数(GEDVI)明显低于扩容无效组[BNP:(158.6±65.8)pg/m L vs(277.4±105.3)pg/m L,P=0.021;GEDVI:(653.4±86.3)m L/m2vs(760.6±102.4)m L/m2,P=0.027]。容量复苏后,扩容有效组和无效组患者间HR、MAP、CVP、SVV、ITBVI、GEDVI、EVLWI比较差异也均无统计学意义(P>0.05),但扩容有效组CO、SVI明显高于扩容无效组[CO:(6.12±1.63)L/min vs(4.89±1.35)L/min,P=0.018;SVI:(56.3±10.4)m L/m2vs(44.6±11.3)m L/m2,P=0.022],而BNP明显低于扩容无效组[(204.7±70.3)pg/m L vs(415.2±124.7)pg/m L,P=0.005]。扩容有效组和无效组患者间△HR、△MAP、△CVP、△ITBVI、△EVLWI差异均无统计学意义(P>0.05),但扩容有效组△SVI、△SVV、△CO、△GEDVI、△BNP均明显高于扩容无效组[△SVI:(19.5±5.7)m L/m2vs(8.9±4.5)m L/m2,P=0.000;△SVV:(2.48±0.82)%vs(0.94±0.25)%,P=0.001;△CO:(2.16±0.78)L/min vs(1.28±0.54)L/min,P=0.001;△GEDV:(86.7±32.4)vs(18.4±8.1)m L/m2,P=0.000],而△BNP明显低于扩容无效组[(55.4±20.6)pg/m L vs(180.7±50.1)pg/m L,P=0.000]。容量复苏前HR、MAP、CVP、SVI、ITBVI、EVLWI与△CO无相关性;而SVV、GEDVI、BNP与△CO呈直线相关(r=0.239,P=0.028;r=-0.846,P=0.021;r=-0.678,P=0.034)。BNP预测容量反应性的ROC曲线下面积(AUC)为0.772,最佳临界值为318.4 pg/m L时敏感度为86.5%,特异度为75.2%;SVV预测容量反应性的AUC为0.726,最佳临界值为11.5%时,敏感度为78.5%,特异度为72.8%;ITBVI预测容量反应性的AUC为0.759,最佳临界值为724.6 m L/m2时,敏感度为83.7%,特异度为73.4%;GEDVI预测容量反应性的AUC为0.827,最佳临界值为675.4 m L/m2时,敏感度为89.2%,特异度为76.5%;BNP、SVV、ITBVI及GEDVI预测容量反应性的AUC、敏感度及特异度均明显高于其他血流动力学指标。结论 BNP可作为预测感染性休克患者容量反应性的无创、简便的指标,可指导临床治疗。

Objective To investigate the value of plasma B-type natriuretic peptide(BNP) as an indicator of cardiac preload and fluid responsiveness in patients with septic shock to guide volume resuscitation. Methods According to the prospective randomized controlled trial(RCT), seventy-eight patients with septic shock in Department of Critical Care Medicine in Hainan Provincial Nongken General Hospital from January 2011 to December 2015 were enrolled in this study, which was given volume resuscitation. Volume resuscitation method was carried out by 6% hydroxyethyl starch(HES) at 300~500 m L/30 min travenously, with the intrathoracic blood volume index(ITBVI) over 850 m L/m2 as an end point of volume resuscition, besides the routine treatrnent. The hemodynamic parameters before and after volume resuscitation were monitored by pulse-indicated continuous cardiac output(Pi CCO). According to the change of cardiac output(△CO) after volume resuscitation, all patients were divided into two groups, responsive group(CO≥15%) and unresponsive group(CO<15%). The hemodynamic parameters, BNP, and the changes in hemodynamic parameters and BNP before and after volume resuscitation were determined. The relationships between hemodynamic parameters, BNP and their changes with △CO were analyzed by the pearson analysis. The role of the hemodynamic parameters and BNP for volume responsiveness prediction was evaluated by receiver-operating characteristic(ROC) curves. Results There were no statistically significant differences in heart rate(HR), mean arterial pressure(MAP), central venous pressure(CVP), cardiac output(CO), stroke volume index(SVI), chest cavity blood volume index(ITBVI) and extravascular lung water index(EVLWI)before volume resuscitation between the responsive group and unresponsive group. However, the stroke volume variation(SVV) in responsive group was significantly higher than that in unresponsive group((13.05±4.75)% vs(9.71±4.38)%, P=0.034), and BNP, global end-diastolic volume index(GEDVI) in responsive group were significantly lower than those in the unresponsive group(BNP:(158.6±65.8) pg/m L vs(277.4±105.3) pg/m L, P=0.021; GEDVI:(653.4±86.3) m L/m2vs(760.6 ± 102.4) m L/m2, P=0.027). There were no statistically significant differences in HR, MAP, CVP, SVV, ITBVI,GEDVI and EVLWI after volume resuscitation between the responsive group and unresponsive group(P>0.05). The CO and SVI of responsive group were significantly higher than those of the unresponsive group(CO:(6.12±1.63) L/min vs(4.89±1.35) L/min, P=0.018; SVI:(56.3±10.4) m L/m2vs(44.6±11.3) m L/m2, P=0.022), and BNP in the responsive group was significantly lower than that in unresponsive group((204.7±70.3) pg/m L vs(415.2±124.7) pg/m L, P=0.005). There were no statistically significant differences in △ HR, △ MAP, △ CVP, △ ITBVI, △ EVLWI between the two groups.However, the △SVI, △SVV, △CO and △GEDVI of responsive group were significantly higher than those of the unresponsive group(△SVI:(19.5±5.7) m L/m2vs(8.9±4.5) m L/m2, P=0.000; △SVV:(2.48±0.82)% vs(0.94±0.25)%, P=0.001; △CO:(2.16±0.78) L/min vs(1.28±0.54) L/min, P=0.001; △GEDVI:(86.7±32.4) m L/m2vs(18.4±8.1) m L/m2, P=0.000), and the △BNP in responsive group was significantly lower than that in unresponsive group((55.4±20.6) pg/m L vs(180.7±50.1) pg/m L, P=0.000). No significant correlation between HR, MAP, CVP, SVI, ITBVI or EVLWI before volume resuscitation and △CO was found. And SVV, GEDVI, BNP before volume resuscitation were correlated with △CO(r=0.239, P=0.028; r=-0.846, P=0.021; r=-0.678, P=0.034). It was shown by ROC curve that the area under ROC curve(AUC) for BNP fluid responsiveness prediction was 0.772, with the cut-off of BNP, sensitivity, specificity of 318.4 pg/m L, 86.5%, 75.2%, respectively. The AUC for SVV fluid responsiveness prediction was 0.726, and the cut-off, sensitivity and specificity were 11.5%, 78.5%, 72.8%, respectively. The AUC for ITBVI fluid responsiveness prediction was 0.759, and the cut-off, sensitivity, specificity were 724.6 m L/m2, 83.7%, 73.4%, respectively. The AUC for GEDVI fluid responsiveness prediction was 0.827, and the cut-off, sensitivity, specificity were 675.4 m L/m2, 89.2%,76.5%, respectively. The AUC, sensitivity and specificity of BNP, SVV, ITBVI and GEDVI for fluid responsiveness were significantly higher than those of other hemodynamic parameters. Conclusion BNP can be used as a noninvasive and convenient index to predict the fluid responsiveness in patients with septic shock, and it can be used to direct clinical practice.