Retinal vascularization begins with the projection of the central hyaloid artery, which regresses around the 16th week of pregnancy.⁽¹⁾ Concomitantly, definitive vascular plexuses form from the optic nerve towards the periphery, controlled by a well-orchestrated gradient of angiogenic factors released by astrocytes in response to fetal physiological hypoxia, and vascularization is completed at the 40th week of pregnancy.⁽¹⁾ In preterm babies, extrauterine high oxygen tension, which can be exacerbated by supplemental oxygenation, downregulates angiogenic factors when the vascularization has not yet been completed, interrupting it (vaso-obliterative phase).⁽¹⁾ As the baby matures, the metabolic demands of the retina grow, leading to a now hypoxic state, which can be intensified by the withdrawal of supplemental oxygenation, and this ultimately triggers a burst of angiogenic factors lacking regulated temporal and spatial gradients, generating pathological neovascularization (vaso-proliferative phase).⁽¹⁾ These events characterize retinopathy of prematurity (ROP), currently one of the main causes of childhood blindness in the world.⁽²⁾ This disease is internationally classified according to three indicators: zones (I to III), stages (1, 2, 3, 4a, 4b and 5) and plus (present or absent).⁽²⁾

Given the large premature population in Brazil, which corresponded to 9.95% of live births in 2019,⁽³⁾ an epidemiological assessment of ROP in the country's health services is necessary. Thus, the present study aimed to evaluate the incidence of ROP and of its outcomes in a reference maternity hospital in northeastern Brazil and to identify associated factors.

In total, 421 standard forms of newborns screened for ROP at MEAC, born between August 2020 and November 2022, were analyzed. Of these, 255 were considered invalid and excluded from the statistical analysis, due to: outcome death (32 forms), outcome hospital transfer (119 forms), incomplete outcome (92 forms) or inconsistency between registered retinal statuses (12 forms).

Table 1 shows the results regarding maternal and neonatal variables, obtained from the 166 valid forms. The following variables were excluded from the analysis due to a lack of sufficient data: occurrence of periventricular hemorrhage (95% of valid forms not completed) and degree of periventricular hemorrhage (98% of valid forms not completed).

Table 2 shows the profile of ROP diagnoses among screened newborns, indicating the stage, zone, and plus classification obtained in the most severe ROP diagnosis during the monitoring period. The incidence of the disease was 40%. No newborn was diagnosed with ROP stage greater than 3. Stage 2 was the most common (70%), followed by stage 1 (23%) and stage 3 (7.6%). No neonate was diagnosed with zone 1 for ROP. Zone 2 was the most incident (94%), with few cases of zone 3 (6.2%). A small portion presented pre-plus (12%) or plus (7.1%). Note that the number of forms with any indication regarding stage (66 forms), zone (65 forms) and plus (42 forms) differed. The mean equivalent gestational age at the time of the most severe ROP diagnosis was 36.39 ± 2.53 weeks. Among those with ROP, 83% had spontaneous regression and 17% had regression after treatment. Among those treated for ROP, all underwent laser photocoagulation.

Table 3 presents the results from the correlation analysis between maternal and neonatal variables with ROP diagnosis and the stage, zone and plus of ROP in the most severe diagnosis. Maternal age (p = 0.046), birth weight (p < 0.001), gestational age at birth (p < 0.001), 1^st minute Apgar score (p = 0.004), resuscitation in the delivery room (p = 0.004), use of surfactant (p < 0.001), use of conventional ventilation (p < 0.001), use of nasal CPAP (p = 0.029), sepsis (p = 0.049), anemia (p = 0.027), blood transfusion (p = 0.028) and apnea (p = 0.008) showed statistically significant associations with ROP diagnosis. Birth weight (p = 0.004) and gestational age at birth (p = 0.010) were associated with the ROP stage at the most severe diagnosis. Birth weight (p = 0.034), gestational age at birth (p = 0.020), and apnea (p = 0.047) were associated with the ROP zone in the most severe diagnosis. Birth weight (p = 0.008) and use of nasal CPAP (p = 0.012) were associated with ROP plus classification in the most severe diagnosis.

Table 4 presents the results from the correlation analysis between the outcome and maternal and neonatal variables. Birth weight (p < 0.001), gestational age at birth (p < 0.001), 1^st minute Apgar score (p = 0.004), resuscitation in the delivery room (p < 0.001), use of surfactant (p < 0.001), use conventional ventilation (p < 0.001), use of nasal CPAP (p = 0.003) and apnea (p = 0.026) were statistically significantly associated with the outcome.

Table 5 shows ROP epidemiological profiles previously reported.^(7-31) Among the 25 studies analyzed, the average incidence of ROP was 30% with a standard deviation of 11%. Among the 21 studies that reported the percentage of babies with ROP who were recommended treatment, the average was 26% with a standard deviation of 16%. In comparison, the present study found an incidence of ROP of 40%, with an incidence of treatment of 17% among those affected. Both values coincide with the scientific literature average within a variation range of one standard deviation.

The slightly higher-than-average incidence of ROP observed may be attributed to the fact that the sample was drawn from a tertiary public maternity hospital, where most women are admitted due to high-risk pregnancies. It is reasonable to assume that neonates in this setting, unlike premature infants born of average-risk pregnancies, accumulate more risk factors, resulting in a higher incidence of ROP. Similarly, the slightly lower-than-expected incidence of ROP treatment could also be explained by sample selection: in a specialized hospital for high-risk pregnancies, the premature death of infants with multiple risk factors is likely more common, preventing them from reaching the stage where more severe cases of ROP would be diagnosed and recommended for treatment. In fact, 32 babies were excluded from the analysis due to death, 22 of whom were never stable enough to undergo ophthalmologic evaluation.

Table 6 shows ROP stage profiles previously reported.^{(7-10,13,15-19,25)} Among the 11 studies analyzed, each stage average incidence and standard deviation was: 37 ± 16% for stage 1, 40 ± 15% for stage 2, 21 ± 12% for stage 3, 1 ± 1 % for stage 4 and 1 ± 2% for stage 5. In comparison, in the present study, stage 2 corresponded to 70% of ROP cases, an incidence considerably higher than the reported average.

Table 6 also displays ROP zone profiles previously reported.^{(7,10,13,15,17,19,29)} Among the 7 studies analyzed, each zone average incidence and standard deviation was: 8 ± 7% for zone 1, 58 ± 32% for zone 2 and 35 ± 33% for zone 3. In comparison, in the present study, zone 2 corresponded to 94% of ROP cases assessed for zone, an incidence higher than the reported average. This result may be explained by sample selection: in a tertiary public maternity hospital specializing in high-risk pregnancies, babies with multiple risk factors are likely to be in such critical condition that they either die before developing zone I ROP or experience significant delays in ophthalmologic evaluation. In the latter, by the time they become stable enough for examination, the disease may have already advanced from zone I to zone II, resulting in a lower number of zone I cases being registered.

Table 6 also shows ROP plus previously reported profiles.^{(14,15,19,31)} Among the 4 studies analyzed, plus was absent in 82% of ROP cases with a standard deviation of 22%. In comparison, in the present study, plus was absent in 81% of the ROP cases assessed for plus, a percentage compatible with the reported average. Only one study considered pre-plus category, reporting an incidence of 22% (Table 6). In comparison, the present study showed a lower incidence of pre-plus, corresponding to 12% of ROP cases assessed for plus.

Unlike other selected articles,^(7,8) maternal age has been identified as a factor associated with ROP. However, the difference between the average age found in mothers of newborns with ROP (28 ± 6 years) was only slightly lower than that of mothers of newborns without ROP (30 ± 7 years).

The association observed between lower birth weight and ROP diagnosis is widely confirmed by other studies.^(7-15,31) Likewise, the correlation found between a lower gestational age at birth and ROP diagnosis is also well described in the scientific literature.^(7-15,31).

The relationship between lower 1^st minute Apgar score and ROP development is confirmed by data reported by other authors.^(7,8) Regarding the positive relationship between resuscitation in the delivery room and ROP, there is also a report corroborating these findings.⁽⁷⁾

Regarding the positive association between use of surfactant and ROP, contrary to the present study, a higher incidence of surfactant use among those without ROP (27%) when compared to those with ROP (14.1%), with a statistical difference (p = 0.01) between groups, has been reported.⁽¹³⁾ Furthermore, the present study found that the use of conventional ventilation was positively associated with ROP diagnosis, while the use of nasal CPAP was negatively associated. Such apparent contradictions could be explained by the dual role of oxygenation in the disease pathophysiology. While post-partum hyperoxia is the main driver of the vaso-obliterative phase of the disease, hypoxia directs the subsequent vaso-proliferative phase.⁽¹⁾ Hence, theoretically, the phase of the disease in which surfactant use or supplemental oxygenation is initiated, as well as its duration and, in the case of oxygen supplementation, its intensity, should affect the outcome regarding ROP. In fact, although oxygen supplementation has been largely considered a risk factor for ROP,^(9,10,31) a recent review⁽²⁾ gathered findings that suggest that lower oxygen saturation ranges during the vaso-obliterative phase, followed by more liberal oxygen saturations (96%) in the vaso-proliferative phase of ROP pathogenesis, could be a strategy that may reduce the severity of the disease.

Regarding the higher incidence of sepsis among newborns with ROP (Table 3), similar results were previously reported.^{(9-11,13,14,16)} The positive association between anemia, apnea, and blood transfusion to a ROP diagnosis (Table 3) is also corroborated by other authors.^(7,10,12,31)

The association between lower birth weight and lower gestational age at birth with a worse ROP zone is expected, since both gestational age at birth less than 30 weeks and birth weight less than 1250 g have been identified as independent risk factors for severe ROP.⁽¹³⁾ The association between more advanced ROP stages with lower gestational ages is also consistent with another report.⁽¹³⁾ However, although birth weight was correlated with ROP stage and ROP plus, the direction of these associations was unexpected, with a lower average birth weight found across intermediate stages and plus classification, respectively stage 2 and pre-plus.

A correlation between apnea and ROP zone was observed, with a higher incidence of apnea across the less severe ROP zones. However, other investigations have demonstrated a higher incidence of apnea (53% versus 36%) among those with ROP progression when compared to those without ROP progression (p = 0.021).⁽⁷⁾

Furthermore, the use of nasal CPAP was associated with the plus classification, with a higher incidence among the non-plus babies. This result reinforces the negative association between the use of nasal CPAP and ROP diagnosis.

Regarding Table 4, the association between lower birth weight and progressively worse outcomes, namely ROP development and a need for treatment, is confirmed by other research. A study found that newborns with ROP requiring treatment were born with statistically lower weights than those who did not require treatment (1,707 ± 597 versus 1,316 ± 346 g; p < 0.0001).⁽¹⁷⁾ Similar results were reported elsewhere.^(9,18)

Regarding the correlation between a lower gestational age at birth and a progressively worse outcome, higher gestational ages among newborns who did not require treatment have already been reported (33w 2d ± 3w versus 30w 4d ± 2w 3d; p = 0.0002).⁽¹⁷⁾ Accordingly, an association between lower gestational age at birth and ROP progression or severity has already been reported.^(7,13) Gestational age at birth has also been identified as an independent risk factor for treatment-requiring ROP.⁽⁹⁾

Regarding the association between lower 1^st minute Apgar score and a worse outcome, statistically lower 1^st minute Apgar scores (p = 0.05) have been reported when comparing babies with ROP stage ≥ 3 (3.6 ± 1.7) and babies with ROP stage ≤ 2 (4.2 ± 1.9), as well as when comparing the latter to babies without ROP (5.1 ± 1.9).⁽¹⁸⁾

Regarding resuscitation in the delivery room, use of surfactant, use of conventional ventilation and apnea, all of them presented higher incidences associated with worse outcomes. However, an association was also observed between higher incidences of nasal CPAP use and a progressively better outcome. These findings might also be explained by the different roles that oxygenation plays during the vaso-obliterative and vaso-proliferative phases of ROP progression. Similarly, another study reported statistically higher incidences of cardiopulmonary resuscitation (64.7% versus 81.8%), use of surfactant (54.4% versus 78.8%), mechanical ventilation (52.9% versus 80.3%) and apnea (36% versus 53%) among infants with ROP progression when compared to those without progression (p < 0.05), while no significant association was found between CPAP use and ROP progression ⁽⁷⁾.

In general, the epidemiological profile found for retinopathy of prematurity and its associated factors in a maternity hospital in northeastern Brazil was compatible with the results found in the literature for several other locations around the world.