The role of PLTs in ROP is not yet fully understood. PLTs are cytoplasmic clumps without nuclei produced by megakaryocytes and have active physiological properties that regulate neovascularization, fibrin formation and deposition (18,19). Studies (19-22) have shown that PLTs locally promote or inhibit angiogenesis. Platelets contain granules that store regulators of angiogenesis, including VEGF and angiogenesis inhibitors (such as endothelial inhibitors), which are released, thus regulating angiogenesis (23). Most studies (24,25) have shown that a reduced PLT count is associated with ROP. Cakir et al (18) discovered that thrombocytopenia at the corrected gestational age of >30 weeks is independently associated with severe ROP that requires treatment. In addition, they found that low PLT count during the neovascularization phase of ROP (phase II) is significantly associated with development of severe ROP.

Şahinoğlu et al (25) also found that a low PLT count is independently associated with severe ROP. Therefore, low PLT count may increase the risk of ROP in preterm infants. Parrozzani et al (23) demonstrated that PLTs act as scavengers in neovascularization by clearing VEGF. A decrease in PLT count results in reduced VEGF elimination, leading to an increase in VEGF and subsequent massive neovascularization. Most studies (21,23,26) have found an association between thrombocytopenia and development or severity of ROP; however, a few studies (7,27) detected no association between low PLT levels and ROP. Choręziak et al (27) found no difference in the early postnatal PLT count between patients with ROP that do and do not require treatment. However, a decrease in PLT count before the diagnosis of ROP is key for the development of ROP. This may be attributed to the immaturity of the blood system at birth. A retrospective study by Özkaya (7) showed no significant difference in PLT between children with ROP requiring treatment and those not requiring treatment. However, the number of children with thrombocytopenia (PLT count #x003C;150x10⁹/l) was higher in the group that needed treatment compared with the group that did not. This may be because the study did not exclude subjects based on number of PLT transfusions or the duration of thrombocytopenia. Lim et al (19) studied mean weekly PLT during the first 6 weeks of life. After adjusting for covariates, there was no significant difference between mean PLT count of preterm infants in the group with ROP and without ROP. Decreased PLT count is not directly linked to ROP, but an effect of other risk factors such as sepsis, blood transfusion and bronchopulmonary dysplasia (19). ROP is influenced by various factors and due to different research and statistical methods, the results can also vary. Recent research (22) examining the impact of stage I and II PLT parameters on ROP has shown no differences in stage I platelet parameters. However, an increase in PLT in stage II with more VEGF release results in increased neovascularization, which contradicts the findings of Parrozzani et al (23). Therefore, the effect of increased or decreased PLT count on ROP is inconclusive, and the pathological mechanisms of PLT count on ROP need further study.

An association has been found between mean PLT volume (MPV) and ROP. MPV reflects PLT size and activity and is a marker of PLT reactivity. During the second stage of ROP, the retina is in a state of hypoxia. The continued increase in VEGF expression stimulates formation of multiple new blood vessels, in which PLTs serve a role in regulating neovascularization and transporting neovascularization factors (28). PLTs in peripheral blood are hyperfunctional, large in size and have a relatively high MPV (28,29). Clinical studies (20,29) have shown a statistically significant difference in MPV levels between children with ROP and controls. Increase in MPV volume raises the risk of ROP by a factor of 1.94(29). MPV levels are associated with diabetic retinopathy, suggesting a link between activated PLTs and proliferative retinopathy (30,31). To the best of our knowledge, there is a lack of research on the association between PLT parameters and ROP; effect of PLTs on VEGF homeostasis and the severity and pathogenesis of ROP requires further investigation.

Reduced Hb levels result in less oxygen being transported. At stage I of ROP, infants with lower Hb levels are unable to meet the increased oxygen demand of the developing retina and retinal vascularization is delayed. At stage II, hypoxia increases VEGF levels, thereby promoting neovascularization (32). Lundgren et al (33) conducted a retrospective study of 227 preterm infants aged ≤28 weeks. Hb levels in the first week of life were significantly lower in infants with ROP requiring treatment than those not requiring treatment. Logistic regression analysis demonstrated the duration of anemia in the first week of life in infants with ROP requiring treatment is an independent risk factor for ROP treatment. Therefore, early prevention of anemia decreases risk of developing ROP. Akyüz et al (34) retrospectively analyzed complete blood count (CBC) parameters in 150 preterm infants; mean corpuscular Hb (MCH) of erythrocytes was the most significant predictor among all parameters. MCH represents the average value of Hb in erythrocytes. The association between MCH and ROP may be associated with the NO pathway. Excess NO can cause vasodilation, capillary leakage and edema, while Hb scavenges NO, resulting in the formation of Hb-NO complexes (34). The aforementioned studies indicate the importance of sufficient Hb in preventing ROP.

Effect of HbA and HbF on ROP after transfusion. HbF serves as the primary oxygen transport protein during fetal development and is involved in transporting oxygen from maternal blood to fetal organs and tissue. In preterm infants, HbF is converted to HbA after birth. HbF has a higher oxygen affinity than HbA. The oxygen dissociation curve of HbF is shifted to the left compared with HbA, thus HbF has a higher oxygen affinity than HbA at any oxygen partial pressure (32). The substitution of HbF for HbA causes the oxygen dissociation curve to shift left, leading to a higher proportion of dissolved oxygen in plasma and an increase in tissue exposure to oxygen (8,35). Studies have shown that low levels of HbF are independently associated with development of ROP and that maintaining higher levels of HbF may protect against ROP (8,32). Therefore, different transfusion regimens may affect outcome of ROP. Children with anemia who receive transfusions of adult red blood cells have a higher incidence of ROP compared with those who receive transfusions of autologous cord blood red cells. In addition, an association is found between the number of blood transfusions and ROP (36). A study suggests that performing ≥3 transfusions before 32 weeks gestation raises risk of ROP (37). The substitution of HbF with HbA during transfusion could contribute to the progression of ROP (38). In a prospective cohort study, Stutchfield et al (39) found a negative association between mean Hb concentration and severity of ROP in postnatal fetuses. They also found that replacing HbF with HbA during transfusion accelerates the development of ROP. This is because transfusing infants with HbA can suddenly deliver a large amount of oxygen to the retina, which downregulates VEGF and leads to stagnation in the development of the retinal vasculature. It has also been suggested that the sudden increase in free iron in blood following the transfusion of adult blood may induce the Fenton reaction, which generates free hydroxyl radicals that damage the retina, thereby inducing retinopathy (40). Moderate blood transfusions are administered to raise Hb levels, which increases oxygen concentration and decreases hypoxic damage to the body. However, excessive adult blood transfusions increase the risk of ROP (39). Therefore, studies in preterm neonates have highlighted that a balance must be found between toxic effects of high-oxygen saturation and the damage caused by hypoxia; during transfusion, it is essential to monitor blood oxygen saturation to prevent anemia and potential damage to the retina from excessive transfusion (16,39). Therefore, prospective research is required to evaluate the predictive value of early anemia detection in preterm infants for ROP.

Influence of glycaemia on ROP. Elevated BG causes biological changes in the retina and most studies (41,42) have found a positive association between BG levels and development of ROP. Mohsen et al (41) conducted a prospective study of 65 children to investigate the association between BG and ROP. They found that maximum and mean BG concentrations were significantly higher in children with than those without ROP. Using logistic regression analysis, they concluded that increase in mean BG concentration in the first week of life is an independent risk factor for developing ROP, which was in accordance with the findings of Vannadil et al (42). In studies (43,44) on the mechanisms by which elevated BG levels affect ROP, it has been found that BG levels are primarily associated with VEGF and IGF-1 and act synergistically. VEGF protein expression in cultured retinal Müller cells is increased at higher glucose concentrations (45), as hyperglycemia can stimulate VEGF production by activating protein kinase C (46). Low serum levels of IGF-1 promote pathological neovascularization in the retina. Studies (43,42) have shown that low levels of IGF-1 inhibit development of normal retinal vasculature, leading to local ischemia and the production of large amounts of VEGF, which leads to hyperproliferative changes in the retinal vasculature. Furthermore, IGF-1 is hypothesized to mitigate insulin resistance. Thus, a deficiency in IGF-1 may result in fluctuating BG levels, potentially contributing to onset of ROP. It is not clear whether hyperglycemia is a clinical manifestation of low IGF-1 levels or the cause of ROP.

Contrary to the findings of Mohsen et al (41) and Vannadil et al (42), Nicolaeva et al (47) found no significant difference in mean BG levels between children with and without ROP and those with spontaneous remission of ROP, ignoring the effect of duration of hyperglycemia on ROP. A meta-analysis by Lei et al (48) revealed a significant association between the duration of hyperglycemia and ROP after adjusting for ORs (odds ratio, but no significant association between average glucose levels and the incidence of ROP. The aforementioned studies therefore suggest that the risk factors for ROP are associated with the duration of hyperglycemia rather than the average glucose levels.

In addition, association between glycated Hb (HbA1c) and ROP has been studied. Movsas and Muthusamy (49) showed that low HbA1c levels are associated with non-proliferative ROP (NP-ROP), while high levels of HbA1c are associated with P-ROP. This may be associated with fetal glucose exposure and metabolic conversion of HbA1c in utero and suggests HbA1c levels in preterm infants as a potential biomarker of ROP. Large-scale prospective studies are needed to confirm the association between levels of HbA1c and ROP.

Risk factors for ROP do not act independently. Instead, multiple factors interact and synergize. For example, it has been shown that hyperglycemia is closely linked to MPV (50). In diabetic patients, vascular endothelial disease accelerates PLT activation and increases peripheral blood MPV, which in turn accelerates the progression of retinopathy (31). The impact of neonatal hyperglycemia on ROP development is still unclear and requires further investigation.

Both maternal systemic inflammation during pregnancy and neonatal inflammatory exposure increase the risk for ROP (51). In addition, maternal inflammatory response during the antenatal period and that of the preterm infant during the postnatal period are associated with the development of ROP. Prenatal inflammation is predominantly histological chorioamnionitis (HCA) and the newborn postnatal inflammation is primarily due to prematurity-related sepsis. HCA is a prevalent inflammatory condition during the perinatal period. HCA is a maternal reaction to infection of the chorionic and placental membranes of the amniotic membrane in the uterus, with bacterial infection as the primary cause of chorioamnionitis (52). The inflammatory response in both the mother and the fetus is the primary risk factor for severe ROP (51). A meta-analysis indicated that maternal chorioamnionitis increases risk of developing ROP (53).

Sepsis is a type of systemic inflammatory response syndrome (SIRS) caused by infection by pathogens (including bacteria, viruses and protozoa). A meta-analysis revealed a substantial association between sepsis and the progression of ROP (54). Early-onset neonatal sepsis exhibits a stronger connection with severe ROP compared with late-onset sepsis. In addition, the majority of cases of early-onset neonatal sepsis are associated with intrauterine infection, indicating a potential association between maternal systemic inflammation and an increased risk of ROP in their offspring (55). Early severe SIRS has been observed in C57BL/6 wild-type mice model to result in abnormal retinal vascular development and increased vascular anastomoses, which are associated with microglia activation (56). Childhood sepsis is associated with the development of ROP and is significantly associated with development of severe ROP (55). A meta-analysis by Wang et al (54) also drew a similar conclusion that sepsis is strongly associated with the degree of ROP at any stage, particularly severe ROP (stages III-V) and sepsis increases the risk of ROP in preterm infants. This may result from damage caused by pathogenic microorganisms and their toxins to the vascular endothelium. This increases the likelihood of leukocytes adhering to vessel walls and forming microthrombi within the small retinal vessels. These microthrombi may obstruct blood flow and cause vascular leakage. Secondly, infected infants may experience pulmonary and respiratory failure as blood pressure drops and hypoxia and changes in blood flow occur (54). In addition, post-sepsis hypotension and changes in oxygen saturation may affect retinal perfusion and worsen retinal ischemia (9). Neutrophil count is the most direct indicator of inflammation. Study has shown that preterm infants with ROP have a higher neutrophil count in their first month than non-ROP preterm infants (57).

Neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) are potential inflammatory markers for diagnosing and prognosing medical conditions, such as acute myocardial infarction and cancer (58,59). Recent studies (10,60) have explored their association with ROP. In a study of CBC in preterm infants 1 week after birth, it was found that NLR is higher in premature infants who develop ROP compared with those who do not and that NLR is independently associated with development of ROP (10). However, a study has not found an association between NLR and the development of ROP but hypothesize that NLR may serve as a risk factor for ROP requiring treatment (60). Contrary to these findings, Ozturk et al (61) found that neither NLR nor PLR are predictive risk factors for treatment in children with ROP. This may be due to the effect of selecting cases of CBC from preterm infants within 24 h of birth and the immature neonatal immune system. Systemic Immunoinflammatory Index (SII) is a relatively new indicator, initially used as a prognostic indicator for hepatocellular carcinoma; however, it is now also used in inflammatory disease (62,63). SII is calculated from lymphocyte, neutrophil and PLT counts, representing a homeostatic balance between inflammatory, immune and thrombotic states (64). SII is calculated as follows: PLT count x neutrophil count/lymphocyte count (65). Akdogan et al (57) first identified SII as an independent predictor of ROP development. They found that preterm infants with ROP have a significantly higher SII than preterm infants without ROP. A recent study (66) proposed that the C-reactive protein/albumin ratio (CAR) may be a marker for the development of ROP. During the first month of life, CAR is significantly higher in preterm infants with ROP than in those without ROP and CAR was also significantly higher in the treated than in the untreated group, suggesting that there may be an association between postnatal inflammation and ROP severity.

In conclusion, inflammation and infection are associated, and maternal chorioamnionitis and sepsis in preterm infants have a strong association with the development of ROP. Neutrophils, NLR, PLR, SII and CAR serve as indicators of inflammatory response in ROP, but the results of NLR in patients with ROP are controversial, which may be associated with time of collection of blood specimens from the children. There are fewer studies on the association between inflammatory factors such as NLR, PLR, SII, CAR and ROP, which needs to be confirmed by a large number of studies.

Effect of inflammatory factors on ROP. The study of inflammatory factors has improved the understanding of the pathogenesis and etiology of ROP (67-70). IL-6 is a proinflammatory cytokine that serves roles in inflammation and immune responses and is a key early indicator of inflammation. IL-8 has multiple functions, including recruiting neutrophils to inflammation sites, cell adhesion, tumor growth, angiogenesis, neuronal protection and brain development (70). Several eye diseases show high levels of IL-8 and IL-6, highlighting the importance of inflammation in these conditions, such as age-related macular degeneration and retinal vein obstruction (71,72). Hellgren et al (73) explored the association between inflammatory factors, insulin-like growth factor I (IGF-I) levels and ROP, demonstrating that inflammatory factors directly or indirectly influence the development of ROP. Elevated postnatal proinflammatory cytokine concentrations are associated with decreased IGF-I levels and ROP since inflammatory factors may inhibit certain components of the IGF-I pathway. Many cytokines share signaling components with IGF-I, such as extracellular signal-regulated kinase 1/2 and mitogen-activated protein kinase. Studies have found that the inflammatory factors IL-6 and IL-8 in amniotic fluid are independently associated with an increased risk of ROP development and progression (67,74). This is consistent with the results of a previous study, which suggested that elevated cord plasma IL-6 levels may be used as an independent predictor of severe ROP and laser treatment (68). Elevated IL-6 is significantly associated with increased risk of developing stage ≥II ROP, suggesting its potential as a biomarker for ROP risk prediction (69). Preterm infants with ROP who require treatment exhibit significantly higher levels of IL-8 compared with those who do not require treatment. However, AUROC (area under the receiver operating characteristic) curves are not effective at identifying serum IL-8 (67,70). The aforementioned studies suggest that the pathophysiological factors that predispose preterm infants to ROP are present prenatally. Therefore, therapeutic strategies to decrease the risk of ROP (specific treatment with antibiotics, anti-inflammatory and/or anti-angiogenic drugs) can be implemented during pregnancy. In conclusion, both prenatal and postnatal inflammatory factors contribute to etiology of ROP.

However, a study (75) did not find that cytokine levels in cord blood (IL-1b, IL-4, IL-6, IL-8, IL-10 and TNF-a) were associated with the risk of ROP, suggesting that measurement of cytokine levels in cord blood samples from preterm infants may be of little value in predicting ROP. Overall, the aforementioned studies highlight the need for monitoring inflammatory markers in preterm infants and future studies should focus on larger, multicenter cohorts and explore the mechanistic pathways between inflammation and ROP.

There are limited studies (11,76) on the effect of lipid metabolism disorders on ROP. Yang et al (11) used mass spectrometry to investigate the link between metabolic changes and the disease. They found significant metabolic disturbances in plasma of the children in the ROP group compared with the non-ROP group, including elevated lipid levels and hyperactivity of lipid metabolism. The retina is rich in lipids and lipid oxidation is a key energy source for the retina. Dyslipidemia can affect pathological angiogenesis in the retina (11). Abnormal activation of lipid metabolic pathways in ROP may lead to lactate accumulation and ketone body production. Dyslipidemia may manifest as elevated serum total cholesterol, low-density lipoprotein and triglyceride levels or decreased serum high-density lipoprotein concentrations (11).

Association between adiponectin (APN) and ROP development has also been investigated (76-78). Adipocytes produce endogenous bioactive protein lipocalin, which regulates lipid and glucose metabolism by promoting fatty acid oxidation and inhibiting lipid synthesis, thus decreasing triglyceride and cholesterol levels in blood. Studies have shown that preterm infants with ROP have lower levels of APN compared with those without ROP (76,77). Lipocalin may inhibit neovascularization by regulating serum levels of TNF-α, IGF-1, ω-3 long-chain polyunsaturated fatty acids (ω-3 LCPUFAs) and VEGF, thereby inhibiting occurrence and development of ROP. Furthermore, preterm birth results in a decreased supply of factors such as IGF-1, ω-3 LCPUFAs and APN from the maternal-placental interface to the fetus. This leads to metabolic disorders, insulin system immaturity and insulin resistance in preterm infants (79). There are few studies (11,76,77) on the effect of lipids and lipocalin levels on ROP in children with ROP and further research is necessary to validate this.