Retinopathy of prematurity (ROP) is a vaso-proliferative disorder of an incomplete retinal vasculature; in the acute phase, it is a time-bound disease and is one of the leading causes of preventable blindness in infants worldwide. ^[1–3] It is a multifactorial disease, and important neonatal risk factors are early gestational age, low birth weight, suboptimal antenatal care, suboptimal neonatal care with supplementation of unblended unmonitored 100 percent oxygen, anemia, respiratory distress, apnea, poor weight gain, and neonatal sepsis. ^[4–5] Maternal risk factors include multiple pregnancy, preeclampsia, maternal diabetes mellitus, oligohydramnios, and maternal infections like chorioamnionitis. ^[4–5] Anemia is a well-known systemic risk factor for retinopathy of prematurity (ROP) in neonates and is caused by an immature hematopoietic system, inadequate erythropoietin production, failure to thrive, iatrogenic blood loss during frequent sampling, poor antenatal care, post-natal infections, maternal malnutrition, and maternal anemia. ^[6–8] Neonatal anemia is defined as hemoglobin that is at least two standard deviations below the mean at a particular gestational and or chronological age secondary to multiple etiologies leading to a reduction in red blood cell mass. ^[6–8] Anemia is a modifiable risk factor, and several studies have shown an association between anemia and ROP development. ^[6–8] Timely correction of anemia with blood transfusion initiates the regression of ROP. ^[7–8] According to the World Health Organization, the severity of anemia in infants can be graded as mild (Hemoglobin 10.0-10.9 g/dl), moderate (Hemoglobin 7.0-9.9 g/dl), and severe (Hemoglobin < 7.0 g/dl).^[9]

Due to multiple factors like small and fragile babies, posterior location of retinopathy, non-dilating pupil, and relatively favourable anatomical and visual outcomes, the intravitreal injection of anti-vascular endothelial growth factor (anti-VEGF) inhibitors has become the primary treatment of choice in aggressive retinopathy of prematurity (A-ROP). ^[10–12] Eyes treated with anti-VEGF need a much longer and frequent follow-up and have a chance of late reactivation and persistent peripheral avascular retina (PAR). ^[10–12] Several studies have investigated baseline fundus changes and fundus fluorescein angiography-based biomarkers to predict progression or regression of ROP post-treatment. ^[13–14] In our previously published report, we have shown the important OCT biomarkers predicting response to intravitreal anti-VEGF treatment in A-ROP eyes.^[11] In this study, we explored fundus and OCT biomarkers in babies with anemia and A-ROP, and their evolution following anemia correction, supplementing the anti-VEGF therapy.

It is a prospective non-randomized case-control study. The institute review board (LEC-BHR-P-09-22-926) approved this study, and written informed consents were taken from parents or guardians of the babies. The study adhered to the tenets of the Declaration of Helsinki. All systemically stable babies examined by the vitreoretinal faculty at the outpatient department (OPD) of our tertiary eye care centre and diagnosed as A-ROP with moderate (blood haemoglobin less than < 10g/dL) to severe (blood haemoglobin less than < 7g/dL) anemia were included.^[9] The babies were prospectively recruited between 1st January 2023 to 31st August 2023 and were followed up for at least 16 weeks. All ROP findings and classifications were according to the revised ICROP3 (2021) recommendations.^[15] As per our institute protocol, a detailed history of gestational age, post-menstrual age, birth weight, and history of blood transfusion was asked. Baseline blood haemoglobin was recorded for all babies with A-ROP during the first OPD visit. Topical mydriatic (tropicamide 1% and phenylephrine 2.5%) was instilled 3 times at an interval of 10 minutes. Topical anaesthetic proparacaine 0.5% was instilled once before inserting the Alphonso paediatric lid speculum. Fundus examination was done with an indirect ophthalmoscope. After informed consent, we performed simultaneous non-contact ultra-widefield (NC-UWF) fundus photo and integrated swept source optical coherence tomography (SS-OCT) with the Optos Silverstone (model name P200TXE, model number A10750) in our retina outpatient department diagnostic setup. We excluded babies who were not systemically stable or babies on oxygen supplementation during the study, previous history of blood transfusion within last 4 weeks, any form of ROP other than A-ROP, ROP stage 3 with elevated extraretinal neovascularization, A-ROP with tractional retinal detachment, babies who had received previous treatment either by anti-VEGF or laser, babies with inadequate details, babies with poor pupillary dilatation or media is too hazy for obtaining interpretable fundus or OCT images, and co-existing non ROP pathologies before retinal intervention for ROP, incomplete follow-ups and lack of parental consent for the study. In this prospective non-randomized case control study, we wanted to evaluate the role of prompt systemic anemia correction with blood transfusion in intravitreal anti-VEGF injected A-ROP eyes versus the babies in whom anemia correction was missed despite reiterating its importance to treating neonatologists, and counseling of caregivers regarding the compliance with prompt systemic anemia correction. All babies were advised to undergo blood transfusion but due to reasons such as logistic concerns in babies coming from rural areas, apprehension of neonatologists and parents regarding risks of transfusion related complications in their judgement, risks of blood transfusion related complications could be higher in rural areas, and current neonatal transfusion guidelines not having ROP as an indication for blood transfusion in babies with moderate anemia (7- 9.9g/dl) were the notable reasons to deny blood transfusion. Some of the important blood transfusion-related complications to expect in neonates were allergic reactions, infections, metabolic imbalances, and fluid overload. ^[6–9]

Supplemental Table 1 shows fundus biomarkers giving clues to underlying severe anemia along with A-ROP. At baseline, Fundus photographs showed a similar proportion of eyes in the two groups with preretinal haemorrhages, deep intraretinal haemorrhages, sclerosed vessels, and Roth spots. Ischemic features on SS-OCT, including hyperreflectivity of inner retinal layers and choroidal thinning of less than 250 microns, were also equally distributed in the two groups. Both groups largely matched the baseline characteristics of retinal fundus findings. Important ones include preretinal hemorrhage seen in 50 percent of eyes, intraretinal hemorrhage in 38 percent of eyes, presence of both superficial preretinal and deep intraretinal hemorrhage (multilayered hemorrhage) in 33 percent of eyes (Fig. 1 and Fig. 2), sclerosed vessels in 17 percent of eyes (Fig. 1 and Fig. 3) and Roth spots in 12 percent of eyes (Fig. 4 and Fig. 5). Ischemic features in OCT, including hyperreflectivity of inner retinal layers, were seen in 45 percent of eyes, and choroidal thinning of less than 250 microns was seen in 38 percent of eyes. Although asymmetric A-ROP was not excluded. By chance, no severe asymmetric presentations of A-ROP eyes were noted in either cases or controls.

Supplemental Table 3, Post intravitreal anti-VEGF injection and prompt blood transfusion there was significant reduction in number and size of preretinal, intraretinal and multilayered hemorrhages in Group 1 (Fig. 2, Fig. 3, and Fig. 5) as compared to Group 2 with p value of 0.024, 0.013 and 0.030 respectively all of which were statistically significant. The average time taken for complete resolution of hemorrhages was 10 weeks in Group 1 and 16 weeks in Group 2. Rapid vascular progression to zone 3 was seen in 70 percent of Group 1 eyes as compared to 25 percent in Group 2 eyes, with a p-value of 0.004, which was statistically significant. Choroidal thickness at the posterior pole was measured within 1000 microns of radius from the foveal centre. Choroidal thickness was measured from the hyperreflective line of the outermost retinal pigment epithelium (RPE) to the innermost hyperreflective line of the choroidoscleral junction. The average choroidal thickness at baseline in Group 1 and Group 2 was 254 microns (range 212–318 microns) and 262 microns (range 210–336 microns), respectively. Both groups showed improvement in average choroidal thickness at 3 weeks post-anti-VEGF injection to 286 and 280 microns, respectively. Reduction in ischemic OCT biomarkers like resolution of hyperreflectivity of inner retinal layers (i.e., inner retinal layers are now delineable from each other) and improvement in choroidal thickness > 250 microns was better in Group 1 when compared to Group 2, although it was not statistically significant.

Supplemental Table 4 shows eyes needing additional treatment with laser due to reactivation of ROP or due to the presence of PAR at 16 weeks post-injection. 12 out of 40 eyes (30 percent) in Group 1 and 32 out of 40 eyes (80 percent) in Group 2 needed additional laser treatment, which was statistically significant with a p-value of 0.001. Six out of 40 eyes in Group 1 showed reactivation, and the average time for reactivation was 13 weeks post-injection. Whereas 18 out of 40 eyes in Group 2 showed reactivation, and the average time for reactivation was 9 weeks post-injection. The remaining 28 eyes in Group 1 and 8 eyes in Group 2 showed complete retinal vascularization till ORA with anti-VEGF monotherapy itself.

Long-term follow-up at 12 months post-anti-VEGF injection is available for 32 babies (14 babies in Group 1 and 18 babies in Group 2). The four eyes of two babies from Group 2 who were previously treated with laser had late reactivation needing repeat anterior laser and or cryotherapy under general anaesthesia due to dispersed vitreous haemorrhage and anterior hyaloid proliferation.

All other babies in both groups remained stable. All the babies were kept under close follow-up with a robust system of calling these babies for follow-up. No babies missed follow-up, and so prompt retreatment with laser could be administered whenever a reactivation was encountered. No eye in either group progressed to retinal detachment or needed surgery.

Supplemental Table 5 shows the Subgroup analysis of babies with moderate and severe anemia at presentation. Severe anemia was noted in 11 babies and moderate anemia in 29 babies. Mean haemoglobin was 6.6gm% in the severe anemia group and 8.7gm% in the moderate anemia group. Eyes with severe anemia had a higher percentage of babies with zone 1 A-ROP (9 out of 11 babies, 82%) when compared to moderate anemia (17 out of 29 babies, 59%). The gestational age and birth weight were comparable between the two.

Our data shows that moderate and severe anemia and its prompt correction by blood transfusion may influence the outcome of anti-VEGF therapy in A-ROP eyes. Clinical manifestations of anemia in newborns include poor feeding, poor weight gain, fast breathing, paleness, neurodevelopmental delays, and chances of being prone to infectious diseases. ^{[9, 19–20]} Therefore, it becomes very important to screen for anemia in all preterm babies and to implement appropriate anemia correction to improve the general health of the baby. ^{[9, 19–20]} Neonatal anemia in literature is sparse as regards ocular manifestations of anemia, especially in preterm babies, and does not provide guidelines for blood transfusion in the presence of severe ocular manifestations.

Ultra-widefield fundus imaging with ultra-widefield swept source OCT helped in detecting possible underlying ischemia. Our study evaluated the role of prompt systemic anemia correction with blood transfusion in intravitreal anti-VEGF injected A-ROP eyes versus the babies in whom anemia correction was missed despite reiterating its importance to the treating neonatologist and counseling of caregivers regarding the compliance with prompt systemic anemia correction.

According to the World Health Organization (WHO), one-third of all women in the reproductive age group are anemic in countries with limited resources and health care infrastructure. ^[21–22] Maternal anemia is a significant risk factor for preterm deliveries and babies with low birth weight. ^[21–22] In our study, 24 out of 40 babies were from a low socio-economic background, referred from public hospitals. Our study did not factor in maternal anemia, though an earlier pilot study by our group has reported less retinal vascularization at birth in babies born to anemic mothers.^[19]

There are no numerical cut-offs for initiation of iron supplements versus blood transfusion for anemia correction, as management is individualized to each neonate after careful assessment of systemic clinical status and etiology for anemia. ^[23] However, mild anemia (Hb < 11g/dl) is usually corrected with oral iron supplements, and it is expected to correct severe anemia (< 7.0 g/dl) with blood transfusion. ^[23] In our study, most of the babies had moderate anemia (14 out of 20 babies in Group 1 and 15 out of 20 babies in Group 2) where the current guidelines require individual decision based on various neonatal factors but do not include ocular manifestations, possibly due to unavailability of routine daily fundus evaluation in every baby in the neonatal centers. The retinal vessels in adults are a part of clinical evaluation to grade and understand the systemic vascular status in various conditions like hypertension, diabetes, anemia, pregnancy-induced hypertension, cardiac and carotid artery diseases, etc., due to easy accessibility through fundoscopy. However, so far, these have not been considered in neonatal care. Our data and earlier studies have shown that anemia is associated with ROP and hence manifests in the neonatal retina. Our study showed that moderate anemia, when corrected with blood transfusion, had favorable outcomes with lower chances of ROP reactivation, lower chances of PAR, and retreatment with laser. A lot of emphasis has been laid on unmonitored, unblended oxygen as a major risk factor for A-ROP. Anemia can significantly contribute to systemic and retinal ischemia because of low hemoglobin and, hence, low oxygen carrying capacity. Its prompt correction with blood transfusion in moderate to severe anemia has not received importance so far, especially from the viewpoint of ocular findings and ROP. Our study of serial fundus photos and OCT biomarkers shows the favorable outcomes post-anemia correction with blood transfusion in A-ROP eyes.

In our study, we captured simultaneous fundus and UWF-SS OCT and assessed them serially in follow-ups to analyze the biomarkers that may be associated with underlying systemic anemia. We found that preretinal hemorrhages, intraretinal hemorrhages, multi-layered hemorrhages, and sclerosed vessels were the four important fundus biomarkers associated with the presence of anemia in A-ROP eyes. Sclerosed vessels have been observed in oxygen-induced retinopathy and anemia-induced retinopathy due to inflammatory damage to the vessel wall.^[24] They are atypical findings in ROP and are getting increasingly recognized in developing countries like India.^[24] Other atypical forms of ROP include hybrid ROP and exudative ROP. ^[24] OCT has revolutionized the evaluation and management of adult vascular diseases and in the assessment of retinal and choroidal ischemia. Hyperreflectivity of inner retinal layers and choroidal thinning are two very important biomarkers suggestive of ischemia. ^{[11, 25–31]} Hyperreflectivity of inner retinal layers is mainly due to cell disorganization with edema, and choroidal thinning is due to ischemia, which in turn results in poor nutrient supply to the outer retina. ^[25–31] In our previously published study, we have shown that OCT biomarkers suggestive of macular ischemia, like hyperreflectivity of inner retinal layers and choroidal thinning, have chance of unfavourable course, reactivation, slower vascular progression, need for retreatment, and need for frequent follow-up post intravitreal anti-VEGF in A-ROP eyes.^[11] These two biomarkers were observed in 45 percent and 38 percent of A-ROP eyes with systemic anemia, respectively, in our study. Group 1 A-ROP eyes treated with intravitreal anti-VEGF with simultaneous correction of systemic anemia showed faster resolution of all forms of hemorrhages and faster vascular progression to the periphery (Fig. 1, Fig. 2, and Fig. 5) when compared to Group 2, and all of which were statistically significant (Supplemental Table 3).

The outcome analysis between the two groups showed that the chances of additional retreatment with laser due to reactivation of ROP or persistence of PAR in Group 2 are much higher than in Group 1, and both of which are statistically significant (Supplemental Table 4). Moreover, Group 2 eyes had an earlier reactivation at the 9th week post-injection compared to Group 1 eyes at the 13th week post-injection, making close follow-up mandatory. We advised blood transfusion in all our babies with A-ROP with anemia. There were no specific worse systemic health outcomes that we noticed in Group 2 after the transfusion was withheld, either by the neonatologist or when the parents did not give consent.

Anemia correction is an associated factor for improving ROP outcomes. ^{[20, 23, 32–36]} Our study findings add to the growing body of evidence supporting anemia correction as an important factor in ROP management. From the published literature, blood transfusion is recommended for those preterm infants with Hb less than 7gm % for systemic concerns irrespective of the ROP status.^[32–36] Our observations showed that A-ROP outcomes with anti-VEGF therapy were better when severe (< 7gm%) and moderate anemia (Hb 7 to 9.9gm%), underwent rapid correction with prompt blood transfusion, in babies who were not on oxygen and were in stable systemic status. This results in a better and quicker regression of the retinopathy. Correction of anemia means better immunity, weight gain, fewer chances of infection, better temperature control, which in turn results in a better response to ophthalmic intervention and treatment. As ROP is a time-bound disease and A-ROP cases worsen in days, these corrections need to happen fast, which is not possible with oral iron supplementation alone. Most of the time, the blood used for transfusion of a preterm infant with anemia is received from an adult or elderly donor. Adult Haemoglobin has less binding capacity with Oxygen compared to Fetal Haemoglobin. ^{[20, 23, 32–36]} As a result, immediately after blood transfusion, there is more release of Oxygen at the tissue level, causing more toxicity to the retinal vasculature and a temporary worsening of the retinopathy. ^{[20, 23, 32–36]} We are aware that blood transfusion in infants, especially when on supplemental oxygen, has been an area of concern. Previous studies have discussed that the multiple blood transfusions with adult haemoglobin in hospital-admitted extremely low gestational age babies may result in disturbances of retinal vasculature due to exposure to higher concentrations of oxygen and reactive oxygen species, leading to a more likely development of ROP or ROP progression, and usually this worsening is temporary. ^[32–36] In these studies, the ongoing oxygen supplementation during NICU admission and unstable systemic course could contribute to the worsening of ROP due to toxicity to the retinal vasculature in addition to the adult blood used for blood transfusion. All the babies in our study were stable, discharged, and not requiring oxygen supplementation. When babies are on oxygen at the time of transfusion, they might need special considerations. In our study, the babies were stable and not on supplemental oxygen, and these adverse events were unlikely to be of concern. Whether the outcomes would be different if they were on oxygen needs a separate study.

The prospective uniform protocol nature of the study, usage of non-contact multimodal imaging biomarkers, and assessment of baseline blood hemoglobin in prognosticating the outcomes of intravitreal anti-VEGF injected A-ROP eyes are the merits of the study. The limitations of the study are that the imaging could be done till 3 weeks of follow-up, as it was difficult to image with the non-contact system in bigger babies, more than 2.5 kilograms. In such babies, however, the fundus drawings/ descriptions with the standard of care indirect ophthalmoscopy were continued, as also contact photography was conducted whenever possible. Outcome analysis was done only for A-ROP eyes injected with Bevacizumab and not for other types of anti-VEGF, as our preferred drug is Bevacizumab only. We included only stable babies, not on oxygen, with moderate and severe anemia needing blood transfusion, but not for milder forms of anemia needing oral iron supplementation. The details on transfusion type, volume, and speed were not collected, and we do not know whether whole blood or only packed cells were transfused; this decision is best judged by the neonatologist. Previous quality of care during neonatal intensive care unit (NICU) admissions, maternal anemia, antibiotic policies to control infections, and oxygen supplementation during previous NICU stays could have effects on outcomes of ROP. There could be other factors, like infections or inflammations, that could be a confounder for both groups. Not many asymmetric A-ROP cases were noted in the study, and this may limit generalizability. The study did not separately analyze the outcomes for specific etiology causing anemia, and the images were analyzed by a single reader masked to whether transfusion was received or not. Treating both eyes of the same patient as independent variables can inflate statistical power, and this is offset to some extent by GEE and confidence intervals.

The study identified a few important multimodal imaging biomarkers suggestive of underlying ischemia in anti-VEGF injected A-ROP eyes with anemia. Important fundus biomarkers include preretinal hemorrhages, intraretinal hemorrhages, multi-layered hemorrhages, and sclerosed vessels. OCT features suggestive of ischemia, like hyperreflectivity of inner retinal layers and choroidal thinning, showed a strong association with anemia. Prompt anemia correction with blood transfusion resulted in faster resolution of hemorrhages and faster vascular progression to the periphery in A-ROP eyes. These eyes showed a lower chance of reactivation and the presence of PAR, needing retreatment with laser. Therefore, anemia serves as an important nutritional biomarker for systemic ischemia, and prompt correction is important for good outcomes in anti-VEGF injected A-ROP eyes. Neonatal transfusion protocols are suggested to take into consideration A-ROP indications for prompt transfusion.