Retinopathy of prematurity (ROP) is a biphasic vaso-proliferative disease of the retina that almost exclusively affects preterm infants. Prematurity, in combination with relative hyperoxia after birth, leads to the arrest of physiological vascular development, involving obliteration and suppression of developing vessels (phase I). As the retina develops further, the metabolic demand exceeds supply, resulting in relative hypoxia and the excessive production of vascular growth factors and neovascularization (phase II) (1). Severe ROP can lead to poor visual outcomes, such as irreversible blindness, if not diagnosed and treated at the appropriate time.

Recent studies have shown that in addition to prematurity and hyperoxia, perinatal infection and inflammation may play a critical role in the pathogenesis of ROP (2). Acute inflammatory conditions of the placenta, such as chorioamnionitis (CA) (maternal host response), funisitis, and chorionic vasculitis (fetal inflammatory response), are caused by chemotactic gradients in the amniotic cavity. The accompanying fetal inflammatory response syndrome may increase the risk of ROP by directly sensitizing the developing retina to oxygen-induced changes in vascular endothelial growth factor (VEGF) levels. Animal studies suggest that perinatal infection and inflammation may sensitize the retina for the development of ROP (2).

The effects of inflammatory mediators and growth factors such as interleukin-1β can significantly increase the activity of hypoxia-inducible factor (HIF-1α) (3). In addition, there may be other inflammatory factors, such as phospholipase-2 and prostaglandins, that affect retinal neovascularization (4).

Clinical data on the association between placental histopathology and retinal outcomes in very preterm infants are conflicting (5, 6). This may partly reflect methodological heterogeneity, specifically the diagnostic criteria for chorioamnionitis (clinical vs. histological) and the gestational and postnatal ages of the study populations. The higher incidence of ROP in larger and more mature babies observed in low- and middle-income countries has been attributed to the increased incidence of infection in these cohorts (7).

In this study, we aimed to characterize the relationship between placental histology and ROP in a cohort of preterm infants < 30 weeks.

Among newborns identified as eligible during the study period (n = 1,670 infants), placental histology and ROP outcome data were available for 878 infants. There were 49 infants with stage 3 or more severe disease and 829 infants with milder disease (stage 0–2). Forty-eight infants needed treatment for ROP, in 31 of whom signs of plus disease were present. Consistent with known risk factors, infants with plus disease were smaller (mean birth weight 669 g vs. 993 g; p < 0.001) than those without. Patient flow and distribution according to ROP stage are depicted in Figure 1.

Descriptive statistics for demographic and clinical variables, by ROP classification, are given in Table 1. The distribution of retinopathy of prematurity classified by histological characteristics is shown in Table 2. A strong association with increasing severity of ROP was observed for oxygen supplementation at 36 weeks (OR: 5.16; 95% CI: 3.81,7.00; p < 0.001), exposure to postnatal steroids (OR: 6.65; 95% CI: 4.02, 11.01; p < 0.001), and receipt of platelet transfusion (OR: 8.21; 95% CI: 4.13,16.34; p < 0.001). In our cohort, more mature gestational age (OR: 0.65; 95% CI: 0.51,1.85; p < 0.01; per 1-week increase) and higher birth weight (OR: 0.62; 95% CI: 0.49,0.79; p < 0.001; per 100 g increase) were associated with decreased odds of severe ROP, and male sex was associated with increased odds (OR: 2.06; 95% CI: 1.08, 3.96; p < 0.05) of severe ROP based on a logistic regression model (with only these three variables included).

After controlling for GA, birth weight, and sex, the presence of maternal chorioamnionitis appeared to show signs of an association with reduced odds of severe ROP: mild chorioamnionitis OR=0.43 (95% CI: 0.17,1.05) and severe chorioamnionitis OR=0.68 (95% CI: 0.29,1.60). While the confidence intervals included 1, the ORs were relatively large in magnitude and somewhat consistent across the levels of maternal chorioamnionitis. After controlling for GA, birth weight, and sex, the presence of fetal chorioamnionitis in the “chorionic plate” appeared to show a significant association with reduced odds (OR: 0.29; 95% CI: 0.11, 0.82) of severe ROP.

When the effects of maternal and fetal chorioamnionitis were examined in the model together (Figure 2), there was some attenuation of the effect. For the most part, the results were stable, with the exception of the OR for maternal severe chorioamnionitis, which became 1.35 (95% CI: 0.34, 5.42); the reduced odds of severe ROP when fetal chorioamnionitis in the “chorionic plate” persisted (OR: 0.26; 95% CI: 0.07, 1.00). The post-hoc combined variable for maternal or fetal chorioamnionitis suggested a reduced risk of severe ROP with just one type of chorioamnionitis (OR: 0.56; 95% CI: 0.21, 1.55) or both maternal and fetal chorioamnionitis together (OR: 0.48; 95% CI: 0.22, 1.07), acknowledging that both confidence intervals included 1. The addition of placental organisms isolated to the model suggested that this variable had little association with severe ROP (OR: 0.91; 95% CI: 0.18, 2.32). The addition of funisitis to the model also suggested that this had little association with severe ROP (OR: 0.92; 95% CI: 0.32, 2.62).

The relationship between placental inflammation and risk for ROP has been studied in both clinical and experimental setups. A systematic review by Mitra et al. did not show any association between CA and ROP (any stage) or with severe ROP (stage 3 or higher) (5). The authors concluded that CA cannot be definitively considered a risk factor for ROP, and further studies should adjust for potential confounding factors and report results by stage to clarify the association with severe ROP. This retrospective study aimed to determine whether there was an association between different stages of histological inflammation and the severity of ROP and to examine common confounding factors. The most recent systematic review by Villamor et al. (2019) stated that histological CA is a risk factor for developing ROP. The review observed a significant positive association between any CA and all stages of ROP. This association was significant for histological but not for clinical CA (6).

In our study, after controlling for GA, birth weight, and sex, the presence of both mild and severe maternal chorioamnionitis appeared to show signs of an association with reduced odds of severe ROP. The OR was 0.43 (95% CI: 0.17, 1.05) for mild chorioamnionitis and 0.68 (95% CI: 0.29, 1.60) for severe chorioamnionitis. Owen et al. recently reported that acute placental inflammation, analyzed by severity, revealed significant inverse associations for both maternal-level (p = 0.01) and infant-level (p = 0.005) inflammation with the presence of ROP disease (14). Park et al. reported, on the basis of a multiple logistic regression analysis (of 85 infants), that amnionitis (OR: 0.120, p = 0.014) and inflammation in Wharton's jelly (odds ratio: 0.124, p = 0.018) were independent protective factors against ROP. It is postulated that with the progression of acute histologic chorioamnionitis, there is a decrease in the level of the intact form of insulin-like growth factor binding protein 1, ultimately affecting IGF-1 bioavailability (15, 16).

A study by Ahn et al. found that the incidence of chorioamnionitis was higher in the APROP infants compared to the infants without APROP (p < 0.001) (17). This is significant, given the higher rate of complications and unfavorable anatomical outcomes following APROP. In our cohort, we could not identify the exact number of APROP cases because of the retrospective nature of the study. However, 14 infants received treatment at a milder stage, which could have been due to the presence of APROP. Kim et al. found a significant association between chronic placental inflammation and severe ROP requiring treatment (adjusted odds ratio: 2.739; p = 0.029) (18).

In a retrospective cohort of 246 infants, a multivariate regression model showed that elevated maternal WBC was associated with ROP progression needing laser treatment (19). The authors could not find any association with histologic chorioamnionitis. A similar finding was noted in our cohort, where maternal or fetal CA did not increase the risk; in contrast, it reduced the risk. In another study involving a retrospective cohort that adjusted for covariates (n = 1217), both maternal and fetal inflammatory markers together posed a significant risk for severe ROP (p = 0.03) (20).

In our cohort of infants, neither early- nor late-onset sepsis had a significant effect on incidence of ROP. The numbers were too small to examine whether sepsis was associated with a higher incidence of plus disease. Wang et al., in a systematic review that had large heterogeneity, found that sepsis increased the risk of ROP in preterm infants (21). Al-Essa et al. found that in preterm infants, sepsis increased the risk of ROP 3.5-fold (22). A higher risk (OR = 6.86) was found by Araz et al. for the development of severe ROP (23).

In this study, the results of the addition of “presence of placental organism” to the model suggests that this variable is not associated with severe ROP (OR: 0.91; 95% CI: 0.18, 2.32). The authors of the ELGAN study (715 infants; GA < 27 weeks) reported that the co-occurrence of bacteria in the placenta and inflammation increased the risk of ROP in Zone 1. If either of these presented alone, it did not increase the risk (24). In our retrospective cohort, information on zone was not available; hence, this observation could not be validated. There was an increasing severity of ROP seen in infants needing postnatal platelet transfusion (OR: 8.21; 95% CI: 4.13, 16.34; p < 0.001) in our cohort. A recent review that included 19 studies concluded that there was an association between platelet deficiency and severe ROP, but this review could not estimate a critical threshold (25).

Our cohort showed a strong association with increasing severity of ROP in infants who needed postnatal steroids (OR: 6.65; 95% CI: 4.02, 11.01; p < 0.001). A similar piece of research involving a retrospective study of 75 infants using a multivariate logistic regression identified “total dosage of systemic steroids” as a risk factor predicting treatment warranting ROP. On receiver operating characteristic (ROC) curve analysis, a cut-off value of 8.95 mg/kg steroid was found to be significant (26). We could not calculate the cumulative dose of steroids in our infants.

Current treatment strategies, both laser and anti-VEGF agents, have their own long-term effects on visual and long-term neurological outcomes. Although anti-VEGF agents have the advantage of being less destructive to the retina, their long-term effects on the developing brain are still being studied. A large body of evidence has emerged to date from the oxygen-induced retinopathy (OIR) models and its molecular pathogenesis, and this does not represent the patient cohort that we see in the clinical world (27). We need more epidemiological data that capture all information, including what happens early in pregnancy and the intrapartum period.

A major strength of this study is the large sample size and the fact that the data included detailed histological assessment of the placenta, including microbiological data. One of the key limitations of this study is its retrospective nature and the exclusion of infants due to a lack of placental histopathology findings. Zone-specific details could have provided more information about the effect of chorioamnionitis on the developing retina.

After controlling for GA, birth weight, and sex, the presence of maternal chorioamnionitis or fetal chorioamnionitis on its own appears to show signs of an association with reduced odds of severe ROP. When the effects of maternal and fetal chorioamnionitis were examined in the model together, there was some attenuation of the effect. The results of the addition of isolated placental organisms or funisitis to the model suggest that these variables are not associated with severe ROP. As reported in the literature, a strong association with increasing severity of ROP was observed in infants who needed oxygen supplementation at 36 weeks and who needed steroids and platelet transfusion in the postnatal period.