早产儿喂养不耐受的风险预测模型分析

Risk Prediction of Feeding Intolerance in Preterm Infants

中文关键词:  早产儿 喂养不耐受 反应范围模型 Logistic回归分析 风险预测模型

英文关键词:Preterm infants Feeding intolerance Reactive scope Logistic regression analysis Risk prediction model

中文摘要:

      目的 在反应范围模型指导下探讨住院早产儿喂养不耐受（FI）发生的风险因素，构建风险预测模型，为有效防治早产儿FI提供实证依据。方法 采用横断面研究，选择2014年8月至2015年1月入住新生儿科的早产儿。在反应范围模型指导下设计临床观察表，从预测性稳态、反应性稳态、稳态超负荷及稳态失败等方面对发生FI及未发生FI者进行分析，通过logistic回归分析建立早产儿FI发生风险预测模型,并检验模型的有效性。结果 ①207例早产儿中，男性125例，女性82例，胎龄27⁺²～37周，平均(33.48±1.66）周；出生体质量830～3 120 g,平均(2 019.55±334.38) g。早产儿FI发生率为33.8％；FI主要发生在开奶72 h内，早期早产儿FI的临床表现以胃潴留为主，晚期早产儿FI则以呕吐为主要临床表现。②单因素分析显示，胎龄、出生体质量、宫内窘迫、宫内感染、氨茶碱应用、母乳喂养及两次大便间隔时间与早产儿FI有关；logistic回归分析显示，胎龄与出生体质量是FI的保护因素，胎龄越大，出生体质量越重，FI发生的风险越低。宫内窘迫、氨茶碱应用、两次大便间隔时间超过3 d是FI发生的高危因素。③经回代预测，回归方程的回代率为92.73%，预测FI的敏感度为97.14%，特异度为88.32% ，准确度为91.30%。 结论 胎龄低、出生体质量低、宫内窘迫、使用氨茶碱及两次大便间隔超过3 d是FI发生的重要风险因素，适用于logistic回归构建预测FI发生风险模型；该模型能够为临床确定FI的高危人群和及时进行干预提供客观可靠的依据。

英文摘要:

      Objective To identify risk factors associated with feeding intolerance (FI) in preterm infants. Methods Preterm infants treated in the neonatal unit of a hospital from August 2014 to January 2015 were recruited in this study. A clinical observation table was developed based on the reactive scope model. Data in relation to predictive homeostasis, reactive homeostasis, homeostatic overload, homeostatic failure and other aspects were collected and compared between those with and without FI.Alogistic regression model was established to determine predictors of FI. Results ①A total of 207 preterm infants were included in the study: 125 male and 82 female. They had an average gestational age of (33.48±1.66) weeks (ranging from 27⁺² to 37 weeks) and an average birth body mass of (2 019.55±334.38) g(ranging from 830 g to 3 120 g).②The incidence of FI was 33.8%. FI in preterm infants often occurred during the period of being fed within 72 h.The main clinical manifestation of FI was gastric retentionin early-preterm infants and emesis in late-preterm infants.② Gestational age, birth body mass, fetal distress, aminophylline application, intrauterine infection, breast milk feeding and interval between stools were associated with FI. Gestational age and birth body mass were found to be significant protectors of FI in the logistic regression model. FI declined with increased gestational age and birth body mass. Fetal distress, aminophylline application, and >3 d interval between stools were found to be significant risks of FI in the logistic regression model.③The prediction model had a 92.73% forecast generation rate of return, with 97.14% sensitivity,88.32%specificity, and 91.30% accuracy.Conclusion Low gestational age, low birth body mass, fetal distress, CM（155.3mm]aminophylline application, and >3 d interval between stools are independent risk factors associate with FI. The prediction model can identify high risk cases of FI.