Inflammation and Retinopathy of Prematurity

The current pathogenetic model

In brief, the current pathogenetic model of ROP postulates that early neonatal exposure of the immature retina to high levels of oxygen reduces the retinal level of pro-angiogenetic growth factors such as vascular endothelial growth factor (VEGF), leading to cessation of vessel growth [12]. Subsequent relative hypoxia then leads to relatively higher VEGF levels and an overgrowth of retinal vasculature. Aspects of vulnerability include genetic components [13] and the relative inability to sustain normal vascularization in the face of adversity (e.g., low levels of VEGF [14] and insulin-like growth factor (IGF) [15].

Infection and ROP

Multiple studies suggest that neonatal infection is a risk factor for ROP. Candida sepsis is associated with a more than fourfold increased risk for ROP (stage 3 or higher) in a multivariable analysis that adjusted for birth weight, gestational age, days on supplemental oxygen, and 5-minute Apgar score [16]. Neonatal sepsis in general (not restricted to Candida) is also associated with ROP [17]. Among more than 15,000 very low birth weight infants from the Israel Neonatal Network, those who had early sepsis were twice as likely to develop severe ROP as their peers who did not have early sepsis [18].

Inflammation and ROP

If hyperoxia and infection are viewed as distal causes of ROP, then oxidative stress and inflammation are the more proximal causes. Oxidative stress, which can be a consequence of inflammation [19], has long been implicated in ROP etiology [20]. Inflammatory processes might interfere with normal retinal vascularization in the most vulnerable retinas. For example, in a murine hyperoxia model of ROP [21], knock-out mice that lack the strongly pro-inflammatory cytokine tumor necrosis factor-alpha (TNFα) appear to be relatively protected [22]. In another study of hyperoxia-induced ROP, omega-3-polyunsaturated fatty acids decreased the size of the avascular area and this protective effect appeared to be mediated, at least in part, via suppression of TNFα [23].

Evidence is just beginning to accumulate that markers of inflammation appear to be associated with ROP in human studies. So far, we know of one report of an association between histologic chorioamnionitis and ROP based on a univariable data analysis [24]. Additional support comes from a more recent and larger observational study of more than 800 extremely low birth weight infants [25]. Over the first three postnatal weeks, the systemic levels of eight cytokines and chemokines were significantly different in infants with ROP compared to controls. For example, the immuno-modulatory interleukin 18 was significantly lower on the first postnatal day, while elevated on days 7, 14, and 21 in infants who later developed severe ROP compared to controls.

Multi-factorial models of disease causation tend to require multiple “hits” [26]. Our data support the hypothesis that multiple hits increase ROP risk. We found that the co-occurrence of immaturity plus pro-inflammatory “hits,” such as clinical chorioamnionitis and neonatal systemic inflammation is associated with an increased risk for ROP [27]. Moreover, we also identified interactions beyond additive (and even multiplicative) patterns among the three “hits” neonatal sepsis, oxygen exposure, and low gestational age (Minghua Chen, personal communication).

Recent evidence supports the notion that poor neonatal weight gain predicts ROP [28]. It will be important to explore to what extent what is attributed to poor neonatal weight gain in ROP etiology should instead be attributed to antecedents of poor weight gain, (e.g., intrauterine inflammation) [29].

Summary and conclusion

In sum, increasing evidence suggests a role for neonatal infection and perinatal inflammation in ROP pathogenesis. The exact mechanism and timing of this inflammatory component of ROP pathogenesis remain to be elucidated, but might deserve the attention of those who wish to prevent ROP and its devastating consequences.