Validation of ROPScore to predict retinopathy of prematurity among very low birth weight preterm infants in a southern Brazilian population

ABSTRACT Purposes: To evaluate the sensitivity, specificity, and cutoff points for the ROPScore, which is based on cumulative risk factors for the prediction of retinopathy of prematurity (ROP), in a population of very low birth weight (BW) preterm infants in southern Brazil. Methods: The medical records of all preterm infants with a very low birth weight ≤1,500 g and/or gestational age ≤32 weeks screened for retinopathy of prematurity in two Brazilian institutions between August 2009 and December 2015 were retrospectively reviewed. ROPScores were calculated using birth weight and gestational age, the use of oxygen therapy with mechanical ventilation, and weight gain proportional to birth weight, as measured at postpartum week six and the need for blood transfusions. Results: The study cohort included 322 infants with a mean birth weight of 1181.8 ± 292.5 g and mean gestational age of 29.5 ± 2.3 weeks. The incidences of any stage of retinopathy of prematurity and severe retinopathy of prematurity were 68.3% and 17%, respectively. ROPScore values ranged from 8.7 to 19.9. The best cutoff point for sensitivity and specificity was 11 for any stage of retinopathy of prematurity and 14.5 for severe retinopathy of prematurity. For any stage of retinopathy of prematurity, the sensitivity and specificity of the ROPScores were 98.6% (95% confidence interval = 97.9%-99.3%) and 35.3% (95% confidence interval= 32.3%-38.3%), with a positive predictive value of 76.6% (95% confidence interval= 74.0%-79.2%) and a negative predictive value of 92.3% (95% confidence interval= 90.6%-94.0%). For severe retinopathy of prematurity, the sensitivity was 100% and specificity was 57.3% (95% confidence interval= 54.2%-60.4%), with positive predictive value of 22% (95% confidence interval= 19.4%-24.6%) and negative predictive value of 100%. The cutoff points correctly identified all infants that developed severe retinopathy of prematurity in this cohort. Conclusions: The ROPScore was useful to identify preterm babies at risk for retinopathy of prematurity. In this population, the ROPScore detected all patients at risk for any stage retinopathy of prematurity and severe retinopathy of prematurity. The ROPScore values in this study were similar to those previously described, thereby successfully validating the ROPScore for early detection of retinopathy of prematurity in very low birth weight preterm infants.

摘要 目的：本研究旨在评估ROPScore——基于累积危险因素的早产儿视网膜病（retinopathy of prematurity, ROP）预测评分工具——在巴西南部极低出生体重早产儿群体中的灵敏度、特异度及最佳截断值。方法：本研究对2009年8月至2015年12月间，巴西两家医疗机构内所有接受早产儿视网膜病筛查的出生体重≤1500g和/或胎龄≤32周的早产儿的病历资料进行回顾性分析。ROPScore的计算指标包括出生体重、胎龄、机械通气氧疗的使用情况、产后第6周测得的体重增长相对于出生体重的比例，以及输血需求。结果：本研究队列共纳入322例早产儿，平均出生体重为1181.8±292.5g，平均胎龄为29.5±2.3周。任意分期早产儿视网膜病及重度早产儿视网膜病的发生率分别为68.3%和17%。ROPScore取值范围为8.7至19.9。针对任意分期早产儿视网膜病，最佳截断值为11；针对重度早产儿视网膜病，最佳截断值为14.5。对于任意分期早产儿视网膜病，ROPScore的灵敏度为98.6%（95%置信区间：97.9%~99.3%），特异度为35.3%（95%置信区间：32.3%~38.3%），阳性预测值为76.6%（95%置信区间：74.0%~79.2%），阴性预测值为92.3%（95%置信区间：90.6%~94.0%）。对于重度早产儿视网膜病，ROPScore的灵敏度为100%，特异度为57.3%（95%置信区间：54.2%~60.4%），阳性预测值为22%（95%置信区间：19.4%~24.6%），阴性预测值为100%。该截断值可正确识别本队列中所有发生重度早产儿视网膜病的患儿。结论：ROPScore可有效识别存在早产儿视网膜病风险的早产儿。在本次研究的群体中，ROPScore可检出所有存在任意分期及重度早产儿视网膜病风险的患儿。本研究中ROPScore的取值与既往文献报道的结果一致，成功验证了ROPScore可用于极低出生体重早产儿的早产儿视网膜病早期筛查。