Zhan Yishan,Zhang Shouhua,Peng Xiaojie,et al.Development and validation of a predictive model for diagnosing acute appendicitis in children excluding anaphylactoid purpura[J].Journal of Clinical Pediatric Surgery,2025,(09):846-852.[doi:10.3760/cma.j.cn101785-202406020-008]
儿童急性阑尾炎排除过敏性紫癜诊断的预测模型开发与验证
- Title:
- Development and validation of a predictive model for diagnosing acute appendicitis in children excluding anaphylactoid purpura
- Keywords:
- Appendicitis; Purpura; Sch?enlein-Henoch; Surgical Procedures; Operative; Child
- 摘要:
- 目的 儿童腹型过敏性紫癜(Henoch-Sch?nlein purpura, HSP)常被误诊为急性阑尾炎(acute appendicitis, AA)，本研究基于大量临床数据，纳入简单、易得的临床指标，旨在开发儿童AA和腹型HSP鉴别诊断的预测模型。方法 回顾性收集江西省儿童医院2010年1月至2020年12月收治的AA及腹型HSP病例完整数据，分为内部及外部验证组；采用单因素分析筛选自分组、R语言进行诊断模型开发与验证、Lasso回归分析筛选预测因子、构建回归模型，K折交叉验证进行内部验证。计算Brier得分、校准截距、校准斜率进行外部验证；因模型构建后是否发热这一指标结局影响较大，另选取无发热患儿按相同方法进行无发热患儿的模型开发与验证，分别绘制受试者操作特征(receiver operating characteristic，ROC)曲线和校准曲线图对两种模型进行评价。结果 共纳入建模组完整数据共5 264例，其中AA 3 375例、腹型HSP 1 889例，外部验证组筛选出903例，包括AA 599例、腹型HSP 304例。鉴别诊断预测模型显示发热(OR=0.22, 95%CI：0.18~0.28)、中性粒细胞比例(OR=0.94, 95%CI：0.94~0.95)、白蛋白(OR=0.86, 95%CI：0.84~0.88)、直接胆红素(OR=0.78, 95%CI：0.74~0.83)、碱性磷酸酶(OR=0.99, 95%CI：0.99~0.99)、C反应蛋白(OR=0.96, 95%CI：0.95~0.96)是诊断腹型HSP的预测因素；基于此模型开发出网页计算器(https://8dc439cf433f6e89f389d65b5116fd53.shinyapps.io/dynnomapp/)。无发热病例3 050例，其中AA 1 384例、腹型HSP 1 666例；验证组无发热病例552例，其中AA 282例、腹型HSP 270例。无发热患儿模型中，中性粒细胞比例(OR=0.95, 95%CI：0.94~0.96)、白蛋白(OR=0.84, 95%CI：0.82~0.87)、直接胆红素(OR=0.80, 95%CI：0.75~0.85)、碱性磷酸酶(OR=0.99, 95%CI：0.99~0.99)、C反应蛋白(OR=0.95, 95%CI：0.95~0.96)、血钾(OR=2.71, 95%CI：2.19~3.38)是诊断腹型HSP的预测因素；基于此模型开发出的网页计算器(https://1af7732f98ae1ac1bc0a023291462db1.shinyapps.io/dynnomapp/)。两组模型内部验证及外部验证的评价指标结果均较满意，建模组内部交叉验证校准后C统计量为0.9450，外部验证ROC曲线下面积为0.8873(95%CI：0.8660~0.9086)，校准截距为-0.3377，校准斜率为0.7325，Brier score为0.1317，无发热病例建模组与验证组ROC曲线下面积分别为0.9164(95%CI：0.9056~0.9272)、0.8472(95%CI：0.8150~0.8794)，C统计量为0.9161，截距为-0.3162，校准斜率为0.7481，Brier score为0.1625。结论 本研究构建的两种鉴别诊断模型具有良好的区分度及校准度，纳入的指标简便、易得，有助于各层次医疗结构对儿童AA排除HSP的诊断，对无发热的病例可选取无发热模型计算器进行鉴别，避免不必要的阑尾切除。
- Abstract:
- Objective In clinical practices, pediatric abdominal type Henoch-Sch?nlein purpura (HSP) is frequently misdiagnosed as acute appendicitis (AA), leading to unnecessary appendectomies.Based upon a substantial amount of clinical data, simple and readily available clinical indicators were used for developing a predictive model for differential diagnosis between pediatric AA and abdominal type HSP. Methods A retrospective collection of complete data were collected on AA and abdominal type HSP cases treated from January 2010 to December 2020.The relevant data were divided into internal and external validation groups.Univariate analysis was utilized for screening independent variables and R language for model development and validation.Lasso regression analysis was utilized for selecting predictive factors and constructing a regression model.And K-fold cross-validation was performed for internal validation.Brier scores, calibration intercepts and calibration slopes were calculated for external validation.Due to a significant impact of fever on the outcomes after modeling, a separate model for afebrile children was developed and validated using the same methods.Both receiver operating characteristic (ROC) and calibration curves were plotted for evaluating both models. Results A total of 5, 264 complete data cases were included in modeling group, comprising AA (n=3, 375) and HSP (n=1, 889).External validation group (n=903) consisted of AA (n=599) and HSP (n=304).The differential diagnosis prediction model indicated that fever (OR=0.22, 95%CI=0.18-0.28), neutrophil ratio (OR=0.94, 95%CI=0.94-0.95), albumin (OR=0.86, 95%CI=0.84-0.88), direct bilirubin (OR=0.78, 95%CI=0.74-0.83), alkaline phosphatase (OR=0.99, 95%CI=0.99-0.99), and C-reactive protein (OR=0.96, 95%CI=0.95-0.96) are predictive factors for diagnosing abdominal type HSP.The model has developed a web calculator: https://8dc439cf433f6e89f389d65b5116fd53.shinyapps.io/dynnomapp/.Among a total of 3050 afebrile cases, there were AA (n=1, 384) and HSP (n=1, 666); validation group included 552 afebrile cases, consisting of 282 AA cases and 270 HSP cases.In afebrile children model, neutrophil ratio (OR=0.95, 95%CI=0.94-0.96), albumin (OR=0.84, 95%CI=0.82-0.87), direct bilirubin (OR=0.80, 95%CI=0.75-0.85), alkaline phosphatase (OR=0.99, 95%CI=0.99-0.99), C-reactive protein (OR=0.95, 95%CI=0.95-0.96) and potassium (OR=2.71, 95%CI=2.19-3.38) were identified as predictive factors for diagnosing abdominal type HSP; the web calculator for this model can be found at: https://1af7732f98ae1ac1bc0a023291462db1.shinyapps.io/dynnomapp/.The evaluation metrics for internal and external validation of both models were satisfactory, with C-statistic for internal cross-validation after calibration in modeling group 0.9450 and area under the ROC curve for external validation 0.8873 (95%CI=0.8660-0.9086) with a calibration intercept of -0.3377, a calibration slope of 0.7325 and a Brier score of 0.1317.For modeling group and validation group of afebrile cases, area under the ROC curve was 0.9164 (95%CI=0.9056-0.9272) and 0.8472 (95%CI=0.8150-0.8794), respectively, with a C-statistic of 0.9161, an intercept of -0.3162, a calibration slope of 0.7481 and a Brier score of 0.1625. Conclusions The above two differential diagnosis models developed here exhibit excellent discriminative and calibration capabilities.The included indicators are simple and readily available, assisting healthcare structures at all levels in excluding HSP in pediatric AA diagnoses.For afebrile cases, afebrile model calculator may be utilized for differentiation, thereby avoiding unnecessary appendectomy.
参考文献/References:
[1] Chen Y, Hu LJ, Wang TT, et al. Clinical mistaken and missed diagnosis of children with abdominal anaphylactoid purpura[J]. Clinical Misdiagnosis & Mistherapy, 2023, 36(5): 11-14. DOI:10.3969/j.issn.1002-3429.2023.05.003. 陈云, 胡立杰, 王婷婷, 等. 儿童腹型过敏性紫癜临床误漏诊分析[J]. 临床误诊误治, 2023, 36(5): 11-14. DOI:10.3969/j.issn.1002-3429.2023.05.003.
[2] Wang LZ, Jin XM. Analysis of Henoch-Schonlein purpura misdiagnosed as acute appendicitis in children: a report of 8 cases[J]. Electron J Clin Med Lit, 2020, 7(22): 40. DOI:10.16281/j.cnki.jocml.2020.22.031. 王立忠, 金秀梅. 儿童腹型过敏性紫癜误诊为急性阑尾炎8例分析[J]. 临床医药文献电子杂志, 2020, 7(22): 40. DOI:10.16281/j.cnki.jocml.2020.22.031.
[3] Liu SX. Analysis of delayed diagnosis in children with abdominal Henoch-Schonlein purpura: a report of 27 cases[J]. Med Sci J Cent South China, 2013, 41(3): 312-313. DOI:10.3969/j.issn.2095-1116.2013.03.027. 刘受祥. 儿童腹型过敏性紫癜延迟诊断27例原因分析[J]. 中南医学科学杂志, 2013, 41(3): 312-313. DOI:10.15972/j.cnki.43-1509/r.2013.03.030.
[4] Jiang ZH. Analysis of misdiagnosis of Henoch-Schonlein purpura: a report of 25 cases[J]. Clinical Misdiagnosis & Mistherapy, 2013, 26(4): 30-31. DOI:10.3969/j.issn.1002-3429.2013.04.011. 蒋志红. 过敏性紫癜25例误诊原因分析[J]. 临床误诊误治, 2013, 26(4): 30-31. DOI:10.3969/j.issn.1002-3429.2013.04.011.
[5] Puig S, H?rmann M, Rebhandl W, et al. US as a primary diagnostic tool in relation to negative appendectomy: six years experience[J]. Radiology, 2003, 226(1): 101-104. DOI:10.1148/radiol.2261011612.
[6] Yang YH, Yu HH, Chiang BL. The diagnosis and classification of Henoch-Sch?nlein purpura: An updated review[J]. Autoimmun Rev, 2014, 13(4/5): 355-358. DOI:10.1016/j.autrev.2014.01.031.
[7] Mizerska-Wasiak M, Skrzypczyk P, Kisiel A, et al. Abdominal symptoms necessitating surgical intervention as the initial presentation of Henoch-Sch?nlein purpura in children-case reports[J]. Pol Merkur Lekarski, 2016, 40(240): 377-379.
[8] Zhang FH, Chen LH, Shang SQ, et al. Atypical purpura location in a pediatric patient with Henoch-Sch?nlein purpura: a case report[J]. Medicine (Baltimore), 2018, 97(48): e13294. DOI:10.1097/MD.0000000000013294.
[9] Fragoso-Maíz C, Ramos-Martínez E, García-Martínez R, et al. Acute appendicitis. A clinicopathological analysis of 200 cases[J]. Rev Gastroenterol Mex, 1993, 58(4): 342-345.
[10] Yang ZB, Feng ZG, Wu DZ, et al. Clinical experience and etiological analysis of acute appendicitis in children[J]. Maternal Child Health Care China, 2021, 36(17): 3994-3997. DOI:10.19829/j.zgfybj.issn.1001-4411.2021.17.030. 杨泽斌, 冯召广, 吴大洲, 等. 儿童急性阑尾炎临床诊治体会和病原学分析[J]. 中国妇幼保健, 2021, 36(17): 3994-3997. DOI:10.19829/j.zgfybj.issn.1001-4411.2021.17.030.
[11] Güng?r A, G?ktu? A, Güneylio?lu MM, et al. Utility of biomarkers in predicting complicated appendicitis: can immature granulocyte percentage and C-reactive protein be used?[J]. Postgrad Med, 2021, 133(7): 817-821. DOI:10.1080/00325481.2021.1948306.
[12] ?zdemir ZC, ?etin N, Kar YD, et al. Hemotologic indices for predicting internal organ involvement in Henoch-Sch?nlein purpura (IgA vasculitis)[J]. J Pediatr Hematol Oncol, 2020, 42(1): e46-e49. DOI:10.1097/MPH.0000000000001571.
[13] Ninchoji T, Kaito H, Nozu K, et al. Treatment strategies for Henoch-Sch?nlein purpura nephritis by histological and clinical severity[J]. Pediatr Nephrol, 2011, 26(4): 563-569. DOI:10.1007/s00467-010-1741-5.
[14] Nevler A, Berger Y, Rabinovitz A, et al. Diagnostic value of serum bilirubin and liver enzyme levels in acute appendicitis[J]. Isr Med Assoc J, 2018, 20(3): 176-181.
[15] Birben B, Akkurt G, Ak?n T, et al. Efficacy of bilirubin values in diagnosing acute appendicitis in patients with normal white blood cell count and predicting complicated appendicitis[J]. Ulus Travma Acil Cerrahi Derg, 2021, 27(1): 50-54. DOI:10.14744/tjtes.2020.91441.
[16] Yang C, Zhou RX, Yang QC, et al. Diagnostic value of serum bilirubin level in acute appendicitis[J]. Chin J Bas Clin Gen Surg, 2014, 21(8): 1013-1015. DOI:10.7507/1007-9424.20140243. 杨晨, 周荣幸, 杨启晨, 等. 血清胆红素升高在急性阑尾炎中的辅助诊断价值[J]. 中国普外基础与临床杂志, 2014, 21(8): 1013-1015. DOI:10.7507/1007-9424.20140243.
[17] Chawla LS, Toma I, Davison D, et al. Acute appendicitis: transcript profiling of blood identifies promising biomarkers and potential underlying processes[J]. BMC Med Genomics, 2016, 9(1): 40. DOI:10.1186/s12920-016-0200-y.
相似文献/References:
[1]饶品德,陈勇.阑尾炎术后急性肾功能衰竭2例报道[J].临床小儿外科杂志,2010,9(06):473.
[2]张振强高群黄河陈晨张燕敏赛恺. 儿童阑尾脓肿的腹腔镜治疗[J].临床小儿外科杂志,2011,10(04):286.
[J].Journal of Clinical Pediatric Surgery,2011,10(09):286.
[3]苏琳 王晓晔 崔华雷. 52例儿童腹型过敏性紫癜诊疗分析[J].临床小儿外科杂志,2012,11(01):28.
[4]周良 侯广军 耿宪杰 张春英 黄华. 小儿过敏性紫癜并发肠套叠的诊断与治疗[J].临床小儿外科杂志,2012,11(04):315.
[J].Journal of Clinical Pediatric Surgery,2012,11(09):315.
[5]赵贤峰 张风华 马新生.急性阑尾炎合并肠梗阻66例诊疗分析[J].临床小儿外科杂志,2012,11(05):391.
[J].Journal of Clinical Pediatric Surgery,2012,11(09):391.
[6]段文强 席红卫. Ghrelin对小儿急性阑尾炎的临床意义探讨[J].临床小儿外科杂志,2014,13(05):422.
[7]张荣鹏 杜国强 孙福涛. 儿童过敏性紫癜合并肠套叠临床分析[J].临床小儿外科杂志,2014,13(06):540.
[8]金晟 吴晔明. 儿童特发性血小板减少性紫癜脾切除术的围手术期处理[J].临床小儿外科杂志,2015,14(05):398.
[9]田俊严 刘丹丹 孙劲松. 新生儿阑尾炎10例诊疗分析[J].临床小儿外科杂志,2015,14(05):417.
[10]雷斌 戚雯琰 徐维. 婴儿急性阑尾炎的临床分析[J].临床小儿外科杂志,2015,14(05):420.
备注/Memo
收稿日期:2024-6-12。
基金项目:2025年江西省卫生健康委科技计划项目青年项目(202510079)
通讯作者:饶小平，Email: 369272877@qq.com