初始剂量静脉用丙种球蛋白治疗无反应川崎病的临床特征和预测指标

Clinical characteristics and prediction of Kawasaki disease unresponsive to initial dose of intravenous immuno. globulin

目的探讨初始剂量静脉用丙种球蛋白治疗无反应川崎病的临床特征和预测指标。方法回顾性分析111例确诊为川崎病患儿的临床资料，根据患儿对初始剂量丙种球蛋白治疗的反应，分为敏感组和无反应组，对比2组患儿的临床表现、实验室检查和心脏超声影像学特征，对2组间存在明显差异的指标行Logistic回归分析，确定独立相关因素，并且以丙种球蛋白疗效分组作为标准，以独立相关因素做受试者工作特征（ROC）曲线，观察初始剂量丙种球蛋白治疗无反应的参考指标。结果1．敏感组90例（81．1％），无反应组21例（18．9％）。2．与敏感组比较，无反应组发生超高热和冠状动脉病变的比例明显增高[超高热：66．7％（14／21例）比34．4％（31／90例），χ2=7．334，P=0．007；冠状动脉病变：47．6％（10／21例）比23．3％（21／90例）,χ2=4．989，P=0．026]。3．丙种球蛋白单次2g/（kg·d）给药与1g/（kg·d）分2次给药相比，患儿发生无反应的比例明显降低[12．5％（9／72例）比30．8％（12／39例），χ2=5．504，P=0．019]，但发生冠状动脉病变无差异[23．6％（17／72例）比30．8％（12／39例）,χ2=0．672，P=0．412]。4．比较2组中性粒细胞比例[（0．72±0．06）比（0．76±0．04），t=-2．84，P=0．005]、血小板[（352．38±42．18）×10^9／L比（373．14±36．93）×10^9／L，t=-2．076，P=0．040]、C反应蛋白[（74．38±12．92）mg／L比（92．05±11．17）mg／L，t=-5．780，P=0．000]等指标，无反应组均明显高于敏感组，但血清清蛋白明显降低[（34．09±3．19）g/L比（31．61±2．03）g／L，t=4．442，P=0．000]。5．多因素Logistic回归分析提示丙种球蛋白治疗无反应的独立危险因素有中性粒细胞比例升高（P=0．018）、C反应蛋白水平升高（P=0．000）和血清清蛋白降低（P=0．040）。6．经ROC曲线下面积计算得出中性粒细胞比例、C反应蛋白和血清清蛋白可作为初始剂量丙种球蛋白治疗无反应有价值的预测指标，临界值分别为0．72、78．5mg／L和33．11g／L。结论当川崎病患儿出现以下情形之一：中性粒细胞比例≥0．72、C反应蛋白≥78．5mg/L或血清清蛋白≤33．11g／L时，应警惕初始剂量丙种球蛋白治疗可能无反应。

Objective To determine the prediction and clinical characteristics of intravenous immunoglobulin （IVIG） treated Kawasaki disease （KD）failure in initial dose. Methods Retrospective analysis was performed with the clinical data of 111 children with KD. The patients were divided into sensitive group and unresponsive group according to initial effect of IVIG. The clinical manifestations,laboratory examination and radiologieal features of the children were compared. Logistic regression analysis was performed in significant different indicators to determine independent correla- tion factors. In order to seek the reference indexes which indicate unresponsive to IVIG, a receiver operating characteris- tic （ROC）curve was made, of which the diagnostic cut -off was nine independent correlation factors while grouping was made according to patients＇ different responses to IVIG. Results （ 1 ） There were 90 cases （ 81.1% ） in effective group and 21 cases （18.9%）in unresponsive group. （2）Compared with the sensitive group ,hyperpyrexia cases[ 66.7% （14/ 21 cases ） vs 34.4 % （31/90 cases） ,χ2 = 7. 334, P = 0.007 ] and the chances of coronary artery lesions [ 47.6 % （ 10/ 21 cases） vs 23.3% （21/90 eases） ,χ2 =4. 989,P =0. 0261 were significantly higher in the unresponsive group. （3） Compared with the children administered twice with gamma globulin, the children of single - dose treatment significantly reduced the unresponsive probability [ 12.5% （9/72 cases） vs 30.8% （ 12/39 cases） ,χ2 = 5. 504, P = 0. 019 ], and there was no difference in the chances of coronary artery lesions[23.6% （ 17/72 cases） vs 30.8% （ 12/39 cases） ,χ2 = 0.672 ,P = 0. 4121. （4）Comparing the sensitive group and the unresponsive group,the percentage of neutrophils count [ （0.72 ± 0.06） vs （0.76± 0.04）, t = - 2. 84, P = 0. 005 ], platelet counts [ （ 352.38 ±2. 18 ） ×10^9/L vs （373.14 ±36.93） × 10^9/L, t = - 2. 076, P = 0. 040 ] and C - reactive protein （CRP） [ （74.38 ± 12.92 ） mg/L vs （92.05± 11.17 ） mg/L,t = -5. 780 ,P = 0. 000 ] were significantly higher in the unresponsive group, but the level of serum albumin [ （ 34.09 ± 3.19 ） g/L vs （31.61 ± 2.03 ） g/L, t = 4.442, P = 0. 000 ] was lower. （ 5 ） Multivariate Logistic regression analysis indicated that the percentage increase of neutrophils count （P = 0.018 ）, CRP（ P = 0. 000） in- crease and serum albumin （P = 0. 040） decrease were independent risk factors for unresponsive treatment with gamma globulin. （6）After the area under the ROC curve was calculated, the percentage of neutrophils count, CRP and serum albumin could be effective predictors to IVIG treatment failure in initial dose, and the critical values were 0. 72, 78.5 mg/L and 33.11 g/L, respectively. Conclusions Treatment with IVIG for the first time may be ineffective in some situations such as the percentage of neutrophils count≥ 0.72, CRP≥ 78.5 mg/L or serum albumin ≤ 33.11 g/L.