Introduction
Retinopathy of prematurity (ROP) is a leading cause of childhood blindness in developed and developing countries worldwide. Approximately 32,300 infants worldwide are diagnosed with irreversible vision impairment due to ROP annually, of which approximately 20,000 become blind or severely visually impaired. Reference Blencowe, Lawn, Vazquez, Fielder and Gilbert1 The main mechanism of this disease includes the arrest of normal retinal and neuronal vascular development that results in aberrant pathological retinal vascularization. Reference Hartnett and Penn2 Laser therapy has been approved to be the gold standard approach in ROP treatment, however it showed some failures especially in severe ROP cases in terms of structural and functional results. 3,Reference Kychenthal, Dorta and Katz4 Later on, studies have demonstrated a clear treatment benefit of intravitreal anti-vascular endothelial growth factor (VEGF) injections in severe ROP cases resulting in prompt regression of the disease. For this purpose, intravitreal bevacizumab (IVB) has been commonly used as an anti-VEGF agent in ROP treatment. Reference Travassos, Teixeira and Ferreira5,Reference Wu, Yeh, Chen, Yang, Lai and Kuo6 In 2011, The Bevacizumab Eliminates the Angiogenic Threat of ROP (BEAT-ROP) study, a prospective, controlled, randomized, and multicenter trial has shown that IVB as monotherapy was significantly useful as compared to laser treatment particularly for the infants with Zone I ROP. Reference Mintz-Hittner, Kennedy and Chuang7
It has been shown that VEGF has a fundamental effect on neurodevelopmental status of the infants, Reference Malik, Vinukonda and Vose8 and concerns have arisen whether systemic suppression of VEGF with IVB may have an influence on long-term neurodevelopment of the premature children. Reference Morin, Luu and Superstein9 Thus far, several studies have evaluated this issue, some indicating no adverse effects of IVB on neurodevelopmental outcomes while others showing higher risk of neurodevelopmental disabilities following IVB. Reference Morin, Luu and Superstein9–Reference Tsai, Yeh, Tsao, Chung, Chang and Lai12 It is suggested that detecting differences in neurodevelopmental outcomes between treatment modalities, evaluating measurable differences, and careful and systematic evaluation of long-term outcomes in these populations. It is also reported that a better understanding of the long-term structural and functional ocular outcomes following anti-VEGF therapy will assist in providing the context and balancing the potential benefits and harms of anti-VEGF therapy in ROP. Reference Tan, Blasco, Lewis, Ostmo, Chiang and Campbell13
In a few ROP studies, there has been discussed visual motor integration (VMI) skills and motor development’s long-term outcomes. Reference Pétursdóttir, Holmström, Larsson and Böhm14,Reference Zimmermann, Schned and Unterasinger15 The impact of ROP on VMI, however, is less well understood. VMI is the ability to perceive visual information, analyze it, and achieve a motor response. It comprises fine and gross motor skills, visual perceptual (VP) abilities, eye–hand coordination, and motor coordination (MC). Reference Beery and Beery16 Being cognitively normal at preschool age but having difficulty in writing, sketching, and throwing or catching a ball abilities that are achieved with VMI skills could be a risk factor for poor school achievement. Reference Dankert, Davies and Gavin17 As far as we know, the differences between the VMI skills of children with ROP according to the types of treatment received were not included in previous studies. The present study aimed to investigate long-term outcomes of VMI and motor development outcomes of premature infants between 2 and 6 years of age who previously had IVB injections and laser treatment. The hypothesis of the study is that there are differences in long-term VMI and motor development outcomes between premature infants who have received prior IVB injection and laser treatment.
Methods
Participants
This cross-sectional study conducted at a single center received approval from the institutional review board of Istanbul Medipol University under the reference number 10840098-604.01.01-E.34158, adhering to the principles of the Declaration of Helsinki. Prior to participation, written and oral informed consent was voluntarily obtained from the parents of each participant. The authors have access to participant information, with measures taken to protect confidentiality both during and after data collection.
The research encompassed children who underwent treatment for ROP from 2012 to 2015 at Zeynep Kamil Maternity and Children’s Diseases Training and Research Hospital, recognized as a leading tertiary referral center for ROP in Turkey. This investigation comprised two distinct groups: the IVB group, comprising children who underwent IVB treatment, and the LPC (laser photocoagulation) group, comprising children treated with LPC. A visual representation of the study’s progression can be found in Figure 1.
Flow chart of the study.
All children were chosen from the hospital’s patient records and contacted via telephone to participate in the study. Those aged between 2 and 6 years with a gestational age (GA) below 34 weeks were eligible for inclusion. Exclusion criteria encompassed children displaying insufficient cooperation during ophthalmologic examinations and neurodevelopmental tests, as well as those with a history of ocular diseases unrelated to ROP or any systemic ailments, including neurological, congenital, and metabolic conditions.
All premature infants in the study underwent standard ROP screening examinations four weeks after birth, adhering to international protocols. 18 Treatment eligibility was determined based on predefined criteria, with options including IVB or LPC. Treatment decisions were guided by established guidelines. 3 Ahead of interventions, parents were briefed on LPC treatment, highlighting its limitations in posterior ROP and potential side effects such as hindering peripheral retinal vascularization and yielding higher refractive outcomes compared to IVB. Reference Kychenthal, Dorta and Katz4 Subsequently, parents were tasked with deciding whether to pursue IVB or LPC treatment. In cases where IVB was chosen, patients and/or guardians were informed of the potential treatment effects and systemic considerations associated with bevacizumab before proceeding with treatment.
Demographic and clinical data of the children including, gender, GA, birth weight (BW), zone and stage of ROP, mode of delivery, multiple pregnancy, and duration of hospitalization in neonatal intensive care unit were recorded. ROP stages were defined according to the International Comitte fort the Classification of Retinopathy of Prematurity (ICROP). 19
Assessments
Visual and refractive assessments
Best corrected visual acuity (BCVA) was measured in all children by using a Snellen chart. Snellen VA was converted to the logarithm of the minimum angle of resolution (logMAR) VA for statistical analysis. The VA results were categorized according to the current World Health Organization visual outcome criteria as follows, Reference Cieza, Keel, Kocur, Mccoy and Mariotti20 mild vision impairment (worse than 6/12 and equal to or better than 6/18), moderate vision impairment (worse than 6/18 and equal to or better than 6/60) and severe vision impairment (worse than 6/60 and equal to or better than 3/60). For the purpose of the present study, normal vision was considered if BCVA was ≥ 6/12 (0.3 logMAR) and abnormal vision was considered if BCVA was < 6/12 (0.3 logMAR).
Automatic refraction was performed with a handheld autorefractometer (HandyRef-K Autorefractometer, Nidek, Gamagori, Japan) in each child after 30 minutes following two instillations of cyclopentolate eye drop. Refractive results were converted into spherical equivalent (SE) values and categorized as emmetropia if SE was (−1 D–1 D), myopia if SE was under −1 D and hypermetropia if SE was above 1 D which was based on previous studies. Reference Quinn, Dobson and Davitt21,Reference Geloneck, Chuang and Clark22
The Beery–Buktenica Developmental Test of Visual Motor Integration
A pediatric physiotherapist with nine years of experience (SKY), who was unaware of the study groups, assessed the neurodevelopmental status of the children. The Beery–Buktenica Developmental Test of Visual Motor Integration (Beery VMI) was employed to gauge the levels of VMI, VP, and MC in the children. Following the administration of the Beery VMI – Short Form test, the Beery VP and MC tests were conducted individually. Utilizing both the total score and the child’s chronological age, the VMI, MC, and VP levels were categorized as “very low, low, below average, above average, high, or very high,” based on normative values outlined in the test’s instruction manual. Reference Beery and Beery16
The Peabody Developmental Motor Scales-2 (PDMS-2)
We utilized the motor development assessment scale to facilitate an early childhood motor development program tailored for children from birth to 5 years old, offering comprehensive evaluation and intervention for both gross and fine motor skills. Reference Rhonda Folio and Rebecca23 Notably, the upper age limit for administering the PDMS-2 assessment is 71 months. Thus, we restricted our study to children aged between 24 and 71 months. The PDMS-2 scores are comprised of three key components: (1) Gross Motor Quotient, encompassing Reflexes, Stationary, Locomotion, and Object Manipulation subtests; (2) Fine Motor Quotient, derived from Grasping and VMI subtests; and (3) Total Quotient, integrating both gross and fine motor subtests. Reference Rhonda Folio and Rebecca23,Reference Wang, Liao and Hsieh24
During the study, under average levels were classified as “abnormal”, average and above average levels were classified as “normal” for both BBDT and PDMS-2 test findings.
Statistical analyses
The sample size of the study was used with G-Power Sample Size Calculator and reference study data within the 95% confidence interval, and Cohen’s d coefficient was taken as the effect size. In the power analysis at the end of the study, the power of the study was calculated to be 83% with a 5% error and a large effect size (0.40). Reference Beligere, Perumalswamy and Tandon25 Statistical Package for Social Sciences-21 (SPSS-21) was used for statistical analysis. Missing values were not inferred. Demographic and clinical characteristics were compared between the groups using Student’s t-test, Chi-square tests, and Fisher’s exact test. For the comparison of visual and refractive outcomes, Fisher’s exact and Chi-square tests were used, respectively. As for neurodevelopmental outcomes, comparison of age equivalents of VMI, VP, and MC between the groups were made by using Student’s t-test. The incidences of BBDT and PDMS-2 test outcomes were compared between the groups by Fisher’s exact test except for VP outcomes which were compared by Chi-square test. Statistically significance was assessed at levels of p < 0.05.
Results
Two hundred ninety four infants with GA < 34 weeks were admitted to our opthalmology department between 2012 and 2015; of these, totally 39 children were accepted in the study. There were 16 children (41%), including eight girls and eight boys in IVB group and 23 children (59%), including 8 girls (34.8%) and 15 boys in LPC group. No significant differences of sex was observed between the groups (p = 0.342). There were no significant differences between the groups in terms of GA and BW (p = 0.060 and p = 0.075, respectively). The disease severity was similar between the groups which was indicated by no significant differences in incidences of ROP zone (p = 0.096) and ROP stage (p = 0.208). The mean duration of hospitalization was significantly higher in IVB-treated children compared to LPC-treated children (p = 0.004). The clinical characteristics of the study population are shown in Table 1.
Clinical characteristics of the study population
SD, standard deviation; ROP, retinopathy of prematurity; IVF, in vitro fertilization; C/S, cesarean section, “+” indicates plus disease.
* Student’s t-test.
** Chi-square test.
***Fisher’s exact test, statistical significance (p < 0.05).
d Cohen’s d.
v Cramer’s V.
Visual acuity and refractive outcomes are summarized in Table 2. The incidence of abnormal visual function was significantly higher in IVB group than in LPC group (p = 0.022). No significant differences were observed regarding refractive status of children in the study. The incidences of emmetropia, hypermetropia, and myopia were similar between the groups (p = 0.173).
Visual and refractive outcomes of the study population
logMAR, logarithm of the minimum angle of resolution.
* Chi-square test.
** Fisher’s exact test, statistical significance (p < 0.05).
v Cramer’s V.
Motor development outcomes are provided in Table 3, Table 4, and Table 5. The mean age at motor development evaluation was found to be earlier in IVB-treated children compared to LPC-treated children (p = 0.044). Referring to BBDT results, mean age equivalents of VMI, VP, and MC did not significantly differ between IVB and LPC groups (p > 0.05). The incidence of abnormal VMI level was significantly higher in IVB group than in LPC group (p = 0.024). Regarding PDMS-2 results, incidences of abnormal gross motor, fine motor, and total motor skills were significantly higher in IVB group than in LPC group (p < 0.05). Regarding PDMS-2 subtest results, incidences of abnormal locomotion, grasping, and visual motor test outcomes were found to be significantly higher in IVB group than in LPC group (p < 0.05).
Comparison of age equivalents of VMI, VP, and MC among groups obtained by Beery–Buktenica Developmental Test
SD, standard deviation; VMI, visual motor integration; VP, visual perception; MC, motor coordination.
* Student’s t-test, statistical significance (p < 0.05).
d Cohen’s d.
The BBDT and PDMS-2 results among groups
BBDT, Beery–Buktenica Developmental Test; PDMS-2, Peabody Developmental Motor Scales-2; VMI, visual motor integration; VP, visual perception; MC, motor coordination.
* Chi-square test.
** Fisher’s exact test, statistical significance (p < 0.05).
v Cramer’s V.
PDMS-2 subtest results among groups
PDMS-2, Peabody Developmental Motor Scales-2.
* Chi-square test.
** Fisher’s exact test, statistical significance (p < 0.05).
v Cramer’s V.
Discussion
The current study shows that VMI and motor development skills were poorer at 2–6 years of age after being treated with IVB compared to premature infants treated with LPC therapy. This is the first study to show the burden of IVB and LPC treatments on long-term VMI skill and motor development in children with ROP.
Over recent years, IVB treatment has become widespread in ROP and studies have shown higher incidence of favorable structural and functional outcomes following IVB compared to LPC treatment. Reference VanderVeen, Melia, Yang, Hutchinson, Wilson and Lambert26 Besides these, investigators demonstrated that IVB significantly reduced plasma VEGF levels with serum drug levels persisting up to two months asserting that IVB can reduce systemic VEGF levels. Reference Kong, Bhatt and Demny27 Then, systemic safety of IVB in preterm children has been questioned as VEGF plays a ciritical role in the developing and adult nervous systems. Reference Rosenstein, Krum and Ruhrberg28 Reduced VEGF expression has been shown to cause decreased neural stem progenitor cell migration which is essential for the continuation of angiogenic and neurogenic coordination. Reference Mani, Khaibullina, Krum and Rosenstein29
Regarding a potential future systemic effect of IVB in preterm infants, studies have been conducted to evaluate neurodevelopmental status of preterm children following IVB for ROP. Lien et al. have found a higher chance of psychomotor impairment in preterm children who received a combination of IVB and laser treatment compared to preterm children who had IVB and laser treatments alone at 2 years of age. Reference Lien, Yu and Hsu10 The authors attributed this result to higher rate of sedation and anesthesia along with a higher incidence of zone I ROP in preterm infants who had combination treatment. Morin et al. have observed lower motor scores in IVB-treated patients than in laser-treated patients, indicating significant neurodevelopmental disabilities in preterm children who received IVB for ROP at 18 months of age. Reference Morin, Luu and Superstein9 But in that study, including children with worse stages of ROP and including children with cerebral palsy have been suggested to impact neurodevelopment outcomes. Kennedy et al. identified no adverse effect of IVB on neurodevelopmental scores compared to laser treatment at the end of a 2-year follow-up assesment. Reference Kennedy and Mintz-Hittner11 In a small case series study, Martinez-Castellanos et al. did not find an apparent deleterious effect of IVB on systemic and neurodevelopmental function in preterm infants by using the Denver Developmental Screening Test II. Reference Martínez-Castellanos, Schwartz and Hernández-Rojas30 Another study evaluated the two-year neurodevelopmental outcomes of patients with ROP who received IVB as adjunctive therapy using the the Bayley Scales of Infant and Toddler Development III (Bayley III). The authors found that IVB did not contribute to systemic complications other than those already present due to prematurity. Reference Araz-Ersan, Kir and Tuncer31 Recently, Fan et al. conducted a prospective case–control study to assess neurodevelopmental outcomes after IVB therapy for ROP and found no significant differences of neurodevelopmental function between premature children who had IVB treatment, premature children with no ROP history and premature children with ROP without treatment between 1 and 3 years of age. Reference Fan, Huang and Huang32 In another study, Chiang et al. did not identify a significant difference of neurodevelopmental function between prematurity without ROP, ROP without treatment, ROP with laser treatment, and ROP with ntravitreal anti-VEGF treatment groups. The results showed that ntravitreal anti-VEGF treatment for ROP was not associated with adverse neurodevelopment in premature infants. Reference Chiang, Chen and Kang33 In the study of Çelik et al., it was stated that children treated with bevacizumab with grade 3 and grade 4 intracranial hemorrhage should be carefully monitored for neurodevelopmental problems. This study concluded that the neurodevelopmental outcomes of anti-VEGF therapy applied to infants with postnatal neurological complications should be followed closely. Reference Celik, Ayranci Sucakli and Kara34
In the present study, we ruled out children with systemic, metabolic, and neurologic disorders initially, thus developmental outcomes were evaluated without the influences of such confounding factors. The Beery VMI and PDMS-2 gold standard methods used in neurodevelopment and psychomotor assessment. We observed significantly higher number of children in IVB group having abnormal VMI skill along with abnormal motor skills compared to children in LPC group. These findings suggest retarded VMI and motor development in IVB-treated children when compared to LPC-treated children at 2 to 6 years of age.
Very preterm infants have a variety of neurobehavioral impairments, such as impaired VMI and VP, however the reported results have a wide range. In a meta-analysis reporting the effect of very preterm birth on VP and VMI skills, it was reported that preterm children born with very low BW had moderate to large effect size impairment in visual-spatial skills but not in visual closure perception. Reference Geldof, van Wassenaer, de Kieviet, Kok and Oosterlaan35 Molloy et al. reported that severe ROP (up to stage 3) was linked to worse results on a variety of neurodevelopmental measures, such as Beery VMI and Bayley III in extremely preterm adolescents. Reference Molloy, Anderson, Anderson and Doyle36 Moreover, Petursdottir et al. showed differences in VMI scores between a control group of term infants and VLBW infants who had previously undergone ROP screening. Reference Pétursdóttir, Holmström, Larsson and Böhm14 Zimmermann et al. reported that Beery VMI scores were significantly lower in preterm infants with ROP stages 2 and 3 compared to infants without ROP, and as a result, the negative effect of ROP on VMI skills in preschool age. Interestingly, preterm infants without ROP in that study had almost normal Beery VMI values. Reference Zimmermann, Schned and Unterasinger15 In this study, it is noteworthy that the average age of evaluation in the Laser group was made earlier than in the Anti-VEGF group. However, Beery VMI and PDMS-2 are norm-referenced assessment scales. Assessment scores were determined to be “normal” or “abnormal” based on age-dependent normative data. It was determined that low BW children with ROP had impaired VMI, VP, and MC skills. VP abilities and the ability to use visual information to guide motor behavior, called VMI have been found to significantly affect a wide range of adaptive abilities, including motor skills such as handwriting, as well as academic achievement in IVB-treated children with ROP when compared to LPC-treated children at 2–6 years of age.
Prematurity has a damaging effect on visual function with its increasing severity. Reference Fielder, Blencowe, O'Connor and Gilbert37 Studies have demonstrated that premature children who received treatment for ROP has limited visual outcomes. Reference Anilkumar, Anandi, Shah, Ganesh and Narendran38,Reference Ng, Connolly, McNamara, Regillo, Vander and Tasman39 In our study, we demonstrated significant differences of visual function between the groups favoring the LPC group. Incidence of abnormal vision was higher in IVB-treated children than in LPC-treated children. These findings suggest better visual performance in children who received LPC treatment in the present cohort. Animal studies have shown that VEGF plays a key role in functioning of adult retinal neuronal cells as well as maintaining retinal pigment epithelium integrity. The results of these studies indicated that VEGF is required for preservation of the visual function. Reference Saint-Geniez, Maharaj and Walshe40,Reference Ford, Saint-Geniez, Walshe, Zahr and D'Amore41 Importance of visual functional status has been emphasized during neurodevelopmental evaluation in preterm children. Better visual performance significantly predicts motor developmental status. Reference Pinello, Manea, Visonà Dalla Pozza, Mazzarolo and Facchin42,Reference van Veen, van Wassenaer-Leemhuis, van Kaam, Oosterlaan and Aarnoudse-Moens43 As we observed abnormal visual function in IVB group, this could have a chance to influence motor developmental test scores in the current study.
It has been shown that extreme prematurity, increased severity of ROP and applying laser treatment in ROP induced myopia. Reference Kaya, Berk and Yaman44 Bevacizumab treatment has been shown to be associated with milder degrees of refractive errors along with a higher chance of having emmetropia. Reference Geloneck, Chuang and Clark22 The effect of IVB on emmetropization has been hypothesized to be related to continuing expression of local growth factors from peripheral retina. Reference Mintz-Hittner and Geloneck45 Our findings regarding refractive status were not consistent with the literature. We could not identify any significant difference of refractive error incidence between IVB and LPC groups at least in this cohort. In our opinion, these results might be due to relatively lower number of subjects in each group and predominantly zone 2 involvement in both IVB and LPC groups.
Premature children with increased severity of ROP have often been sicker newborns during neonatal period suggesting that these children tend to have delayed neurodevelopmental outcomes compared to their age equivalents. Reference Goyen, Todd, Veddovi, Wright, Flaherty and Kennedy46 In our study, although we observed no differences of ROP zone and stage between the groups, we identified significantly longer duration of hospitalization in IVB-treated children. This finding was in contrast with that of in the studies by Kennedy et al. Reference Kennedy and Mintz-Hittner11 which found shorter hospital stay in IVB group than laser group and Chen et al. Reference Chen, Schachar and Moshfeghi47 which found no significant difference of length of hospital stay between IVB- and laser-treated children during neurodevelopmental assessment. As we could not reach detailed patient records, our finding on hospitalization time might suggest a trend for IVB-treated children to have multiple comorbide factors such as sepsis, intraventricular hemorrhage, patent ductus arteriosus, increased time of mechanical ventilation and oxygen fluctuation that may have an association with impaired motor developmental scores. Reference Stoll, Hansen and Bell48 One may find the difference in hospital stays across the present sample suggestive that at least part of the developmental outcome differences observed are likely related to illness level differences between the groups.
The present study has some limitations, including small number of subjects, nonrandomizing the groups and lack of knowledge of multiple comorbid factors during the postnatal hospital period. In addition, outcome times during neurodevelopmental assessment vary between 2 and 6 years when there is likely to have been a significant improvement in function. Furthermore, most reports have used the Bayley III to evaluate neurodevelopmental status in premature children. As we utilized the Beery VMI and PDMS-2 tests, we were not able to identify cognitive, language, social-emotional, and adaptive behavior skills of the children in our study. Nonetheless, these tests have been shown to have higher predictivity and reliability to measure visual motor development and to better reveal motor skills in premature children. Reference Beery and Beery16,Reference Wang, Liao and Hsieh24,Reference Sutton, Barchard and Bello49
Conclusion
In conclusion, significantly lower VMI and motor development features was demonstrated in IVB group than in LPC group between children 2 and 6 years of age in our study. In this study, the hypothesis that there is a difference in long-term VMI and motor development outcomes between premature infants who have previously received IVB injection and laser therapy is confirmed. Cases treated with IVB had chance to be sicker infants with longer hospital stays, it may therefore not be treatment that was the cause of the development issues but the general level of sickness. Notwithstanding, our results suggest the relevance of IVB injection in impaired motor developmental outcomes in premature children which should be bear in mind. Depending upon visual functional outcomes in our study, we recommend a comprehensive evaluation for IVB-treated children with regard to early physiotherapeutic programs to enhance visual and motor development. Prospective, randomized, case–control, and larger series studies with longer follow-up are needed in order to better ascertain the effect of IVB on neurodevelopment.
Acknowledgments
We thank children, their families, and Asena Ayca Ozdemir who assisted with the statistical analysis.
Financial support
None.
Competing interests
None.
