青少年大骨节病诊断指标筛选和模型的构建

Screening and model construction of the clinical diagnostic indictors for Kashin-Beck disease in adolescents

目的了解青少年大骨节病临床指标与大骨节病关系及在临床诊断中的意义，构建青少年大骨节病诊断模型。方法在陕西省采取整群随机抽样的方法抽取6个病区县和1个非病区县的青少年，共计2248人。分析指标包括性别、年龄、大骨节病临床和影像指标共计32项。指标组间分布计量资料用t检验和方差分析，计数资料用χ^2检验，采用多分类有序反应变量Logistic逐步回归分析进行模型构建。结果青少年大骨节病患病率随年龄变化有增高趋势，组间比较除年龄因素（P〉0．05）外，踝关节疼痛、膝关节疼痛、腕关节活动障碍等5项（P〈0．05）和指末弯、肘关节活动障碍、指畸形等9项（P〈0．01）比较差异有统计学意义。指标中16项与大骨节病严重程度显著相关（P〈0．01），构建4个具有统计学意义的累积logit模型（-2logL、Score、Wald、χ^2检验，P〈0．01）。结论建立了以全身受累关节为表现的指标诊断模型，依据指标贡献大小，对于临床诊断具有重要的辅助作用。

Objective To understand the relationship between clinical manifestations and Kashin-Beck disease（KBD） and their contribution to diagnosis of KBD and to construct the diagnosis model for KBD in adolescents. Methods A total of 2248 subjects under the age of 18 were collected from 6 KBD endemic and 1 non-KBD areas of the Shaanxi province in China. Analysis of 32 indicators, including gender, age, and KBD elinieal indicators. Indicators of the distribution of measurement data between the two groups using t test and analysis of variance, χ^2 test with count data, multi-eategory ordered response variables Logistic regression analysis for model building. Results It showed the KBD prevalence rate in adoleseent had an increasing tendency with age. Analysis of indicators between the two groups, in addition to the age factor （P 〈 0.05 ）, the difference of ankle pain, knee pain, wrist movement disorder and other 5 indicators （P 〈 0.05） and the last bend, elbow movement disorder, syndaetyly and other 9 indicators（P 〈 0.01 ） were statistically significant. Sixteen elinieal and radiographic features in the clinical manifestations were significantly related with the clinical severity grading with KBD （P 〈 0.01 ）. Four models on the diagnostic indictors were constructed by cumulative logit model for adolescent KBD （-2logL, Score, Wald χ^2 test, P 〈 0.01 ）. Conclusions The establishment of the diagnostic model based on their contribution of the joint involvement in systemic performance-related indicators has an important role for clinical diagnosis of KBD.