Shah Sarthak V, Brant Arthur R, Ahmed Akwasi, Wireko Brobby Naana A, Zhao Cindy S, Ahya Ashna, Tabin Geoffrey C, Moshfeghi Darius M
Byers Eye Institute, Horngren Family Vitreoretinal Center, Dept. of Ophthalmology, Stanford University School of Medicine, Palo Alto, California, United States.
Komfo Anokye Teaching Hospital Department of Ophthalmology, Kumasi, Ghana.
Invest Ophthalmol Vis Sci. 2025 Aug 1;66(11):17. doi: 10.1167/iovs.66.11.17.
Accurate identification of retinal Zone I in retinopathy of prematurity (ROP) is critical for treatment decisions and prognosis. Current definitions rely on identifying the macular center, limited by absence of the foveal light reflex (FLR) early in screening. Understanding factors influencing FLR development could improve zone localization and guide nutritional interventions. We sought to determine whether absolute infant weight gain independently predicts FLR development in premature infants, beyond postmenstrual age (PMA).
Prospective observational cohort study conducted from November 2024 to March 2025 within a multicenter telemedicine-based ROP screening program (GHANAROP). Premature infants meeting ROP screening criteria, screened before 37 weeks PMA, who subsequently developed FLR identifiable on wide-field digital retinal imaging were included. Logistic regression assessed associations between FLR presence, PMA, and infant weight. Multivariate logistic regression with elastic net regularization evaluated combined predictive value of PMA and weight.
Among 318 eyes, FLR was first identified at mean PMA of 35 0/7 weeks (low risk ROP) and 35 2/7 weeks (more than low risk ROP), and mean weight of 1723g (low risk) and 1565g (more than low risk). Univariate modeling found that each additional week of PMA increased odds of FLR presence by 44.6% (OR = 1.446; 95% CI, 1.371-1.524) and that each 200g increase in weight increased odds by 23.1% (OR = 1.231; 95% CI, 1.113-1.360). Multivariate modeling demonstrated PMA and weight as predictors of FLR development, achieving 84.8% predictive accuracy (Nagelkerke R² = 0.2783).
Absolute infant weight independently predicts FLR development beyond PMA alone, highlighting opportunities for nutritional interventions to accelerate foveal maturation and potentially enhance visual outcomes.
准确识别早产儿视网膜病变(ROP)中的视网膜I区对于治疗决策和预后至关重要。当前的定义依赖于识别黄斑中心,但在筛查早期受限于中央凹光反射(FLR)的缺失。了解影响FLR发育的因素可以改善区域定位并指导营养干预。我们试图确定婴儿绝对体重增加是否独立于孕龄(PMA)预测早产儿FLR的发育。
2024年11月至2025年3月在一个基于多中心远程医疗的ROP筛查项目(GHANAROP)中进行前瞻性观察队列研究。纳入符合ROP筛查标准、在孕37周前进行筛查且随后在广角数字视网膜成像上可识别出FLR的早产儿。逻辑回归评估FLR存在、PMA和婴儿体重之间的关联。采用弹性网正则化的多变量逻辑回归评估PMA和体重的联合预测价值。
在318只眼中,FLR首次在平均孕龄35 0/7周(低风险ROP)和35 2/7周(高于低风险ROP)时被识别,平均体重分别为1723g(低风险)和1565g(高于低风险)。单变量模型发现,PMA每增加一周,FLR存在的几率增加44.6%(OR = 1.446;95% CI,1.371 - 1.524),体重每增加200g,几率增加23.1%(OR = 1.231;95% CI,1.113 - 1.360)。多变量模型表明PMA和体重是FLR发育的预测因素,预测准确率达到84.8%(Nagelkerke R² = 0.2783)。
婴儿绝对体重独立于单独的PMA预测FLR的发育,突出了营养干预加速中央凹成熟并潜在改善视觉结局的机会。