This study demonstrated that systemic apelin-13 and VEGF may play a role in assisting the diagnosis of ROP with a relatively high sensitivity and specificity and that systemic apelin-13 concentrations are negatively correlated with systemic VEGF concentrations in infants with ROP.
Improvements in our understanding of the molecular basis of diseases and advances in technology have fuelled the search for novel biomarkers for many diseases. Biomarkers can potentially improve our ability to identify, manage, or prevent a wide range of conditions that jeopardise patient health. Biomarkers are mostly used to identify the presence of a disease, selectively treat the population at greatest risk, and provide a low risk–benefit ratio .
Preterm infants usually require supplemental oxygen during post-natal treatment. Exposure to relatively high oxygen levels delays the formation of retinal blood vessels, causing vascular occlusion. This stage usually occurs at 30–32 weeks of gestation. When blood vessels resume growth, new abnormal blood vessels are formed by the involvement of multiple cytokines . Animal models of oxygen-induced retinopathy have shown that VEGF is a key player in the pathogenesis of ROP . An in vivo study on ROP has found a retinal VEGF pattern consistent with that detected in in vitro studies . Concurrently, animal studies on retinal endothelial cell lines have demonstrated that apelin-13 has angiogenic activity [18, 19]. Many animal studies and a small number of human studies have demonstrated that apelin-13 is associated with the occurrence of ROP [7,8,9]. Thus, systemic apelin-13 and VEGF may be meaningful biomarkers to aid in the diagnosis of ROP [9, 20].
Currently, there is a consensus on the inverse correlation between the incidence of ROP and gestational age and birth weight. Given the different manifestations of ROP and various external factors, neonatal birth conditions and peri-natal treatment strategies, such as oxygen supply and blood transfusion strategies, must be given appropriate consideration in the neonatal intensive care unit [21, 22]. In our study, the selected control infants were matched in terms of gestational age and birth weight, and after comparison, sex, number of multiple births, Apgar score, oxygen supplementation, and blood transfusion were similar between the ROP and non-ROP groups. This enabled us to avoid the bias that might have been caused by the above factors in the comparison of apelin-13 and VEGF cytokines.
Our study found that preterm infants with a gestational age of < 32 weeks and birth weight of < 1500 g had higher plasma VEGF levels and lower plasma apelin-13 levels in infants with ROP than those in infants without ROP at 4–6 weeks after birth; and infants with severe ROP had higher plasma VEGF levels and lower plasma apelin-13 levels than infants with wild ROP. The results of several other in vivo studies on venous blood VEGF levels [4, 20] are consistent with the results of this study, particularly in terms of gestational age, birth weight, and sample evaluation time. However, our results are different from those reported by Pieh et al.  and Kandasamy et al. ; this may due to the different gestational ages, birth weights, and sample evaluation times of preterm infants in these studies. Differences in gestational age at birth, birth weight, and oxygen supplementation time after birth may affect cytokine levels. In terms of research on blood apelin-13 levels, both Ali et al.  and Cekmez et al.  observed significantly lower blood apelin-13 levels in the ROP group than in the control group, which is consistent with the results of our study. These two studies were also single-center studies with small-sample sizes; therefore, further research is needed. Feng et al. , also from our institution, reported results that are opposite to ours. They found that the VEGF levels in the ROP group were lower than those in the control group, while the apelin-13 levels were higher in the ROP group than in the control group. The reason may be that their study population included all pre-mature babies as they did not specify that the gestational age should be < 32 weeks. Another reason for the difference in the results may be due to their specimen sampling time, which was > 6 weeks after birth until approximately 3 months, as it is generally known that cytokine levels will change with time after birth . The time point used for ROP screening in our study—4–6 weeks post-partum—maybe more meaningful for detecting differences in cytokine levels and, therefore, for predicting the disease. Data on the systemic levels of VEGF and apelin-13 related to ROP have been inconclusive; thus, our results provide evidence that plasma apelin-13 and VEGF levels measured at 4–6 weeks of age may play a role in the diagnosis of ROP.
In terms of the relationship between apelin-13 and VEGF, Huang et al.  demonstrated that apelin markedly upregulated VEGF protein expression at all selected time points after middle cerebral artery occlusion, whereas Uribesalgo et al.  showed that apelin, at least in part, suppressed the VEGF transcriptome in tumour vessels. In addition, Hou et al.  performed experiments in mice to confirm that apelin could effectively promote angiogenesis under hypoxic–ischaemic conditions, which is related to the up-regulation of VEGF. Likewise, the relationship between systemic apelin and VEGF levels in neonates with ROP is inconclusive. Zhang et al.  reported that there was no correlation between apelin and VEGF expression in neonates with ROP, whereas Feng et al.  found that the plasma levels of apelin and VEGF were negatively correlated in children with ROP. In this study, we found a negative correlation between plasma apelin-13 and VEGF levels, not only in ROP group but also in non-ROP group, which suggesting that apelin and VEGF may mutually regulate each other , but the mechanism is still unclear. Since the negative correlation may be related to ROP or preterm birth, further studies with large samples are needed to confirm this correlation.
In recent years, there has been an increasing interest in determining the related pathogenic mechanisms of ROP  and related biomarkers such as apelin-13 and VEGF, but there is insufficient conclusive evidence to support their usefulness; our research results provide some evidence. The ROC curve was used to define the best apelin-13 cut-off value for our patients, which was 119.6 pg/mL, with a sensitivity of 84.8%, specificity of 63.6%, and diagnostic accuracy of 80.4%. Another ROC curve was used to define the best VEGF cut-off value for our patients, which was 84.3 pg/mL, with a sensitivity of 84.8%, specificity of 66.7%, and diagnostic accuracy of 81.0%. This finding is in agreement with that reported by Ali et al.  and suggests a role of plasma apelin-13 and VEGF in the diagnosis of ROP.
This study has some limitations, including its single-center study design and relatively small number of included cases. Future multi-center studies with large samples are needed to confirm the clinical relationship between various biomarkers, such as apelin-13 and VEGF, in patients with ROP to provide innovative ideas for the screening of ROP.