Background: Preterm infants are at risk for neurodevelopmental deficits. An association between retinopathy of prematurity (ROP) and impaired cognitive outcome has already been described. However, less is known about the impact of ROP on visual motor integration (VMI), which is a prerequisite not only for fine motor abilities but also for further school skills. Therefore, the aim of this study was to retrospectively investigate the impact of ROP on VMI at preschool age. Methods: The study was conducted at the Medical University of Vienna, including patients born between January 2009 and December 2014 with a gestational age of less than 30 weeks and/or a birth weight of less than 1,500 g. VMI was determined by Beery-Buktenica Developmental Test of Visual Motor Integration (Beery VMI) at the age of 5 years. Results: Out of 1,365 patients, 353 met inclusion criteria for this study. Two hundred sixteen of them had no ROP, while 137 had ROP (stage 1: n = 23, stage 2: n = 74, stage 3: n = 40). Mean value of the Beery VMI score was significantly lower in the ROP group compared to the No-ROP group (90 ± 16 vs. 99 ± 14; p < 0.01). By correcting for other important medical conditions, ROP still had a significant impact on Beery VMI score (p < 0.01). Particularly, lower scores were found for stage 2 (p < 0.01) and stage 3 (p < 0.01). Conclusion: Beery VMI scores were significantly lower in preterm infants with ROP stage 2 and 3 than in infants without ROP. This study shows the negative impact of ROP on VMI skills at preschool age, even after adjustment for key demographic and medical characteristics.
Advancements in ante- and neonatal care, monitoring, and ventilatory support have permitted infants of lower birth weights and younger gestational ages (GAs) to survive . Complementary to a decreasing mortality rate of preterm infants with a GA below 32 weeks and very low birth weight (VLBW), also a decrease in total number of morbidities has been observed [2, 3]. Nonetheless, many of these morbidities, including the retinopathy of prematurity (ROP), still have a major impact on infants’ outcomes [2, 4, 5].
ROP is an ocular disease of the developing retina of premature infants that affects the structure and function of the immature vascular system of the eye. It is considered as one of the most common causes for blindness and visual impairment in premature infants. Globally, 32,300 children were estimated to be visually impaired in 2010 as a result of ROP . The main characteristics of this disease range from visible vascular abnormalities up to retinal detachment appearing mostly in the second and third month after birth . Because of the increased number of premature birth, the number of ROP is rising worldwide and especially in developing countries, where neonatal care is improving but optimal oxygen administration is still a major challenge .
In a few studies of both ROP and neurodevelopmental outcome, a link between cognitive performance and ROP has been discussed [9, 10]. However, less is known about the influence of ROP on visual motor integration (VMI). VMI is the capacity to perceive visual information, process it, and coordinate a motor response . It includes visual perception skills, eye-hand coordination, gross motor coordination, and fine motor coordination . Basic child activities such as handwriting, drawing, and throwing or catching a ball need VMI skills, and these skills are especially crucial after school entry . Having a normal cognitive development at preschool age but presenting weakness in VMI might be a risk factor for adverse performance at school.
Considering this assumption, the main aim of this study was to investigate the impact of ROP in its different stages on VMI at preschool age. Additionally, information on cognitive development of the population was collected.
This is a retrospective data analysis including inborn infants born at the Medical University of Vienna (MUV) between January 2009 and December 2014. The study was done in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the MUV (EC No. 1554/2020). Given the retrospective nature of the study, informed consent was not required.
Children born after 29+6 weeks GA and/or weighing more than 1,500 g, presenting major birth defects and/or malformations, were excluded from this study. Clinical data, including ROP, were extracted from the electronic patient documentation system and defined according to the definitions used in the Vermont Oxford Neonatal Network. Follow-up data were collected by the follow-up clinic of the Department of Pediatrics and Adolescent Medicine. All infants admitted to our unit, born before the 32nd week of gestation, are further assessed for neurodevelopment in our follow-up clinic by paediatricians and a certified clinical psychologist, following a standardized assessment program. According to the above-mentioned programme, the paediatricians are responsible for the general health examination of the infant. Vision problems are documented in a follow-up sheet at the age of 4 years after an external ophthalmological examination, which is routinely documented in the Mother-Child-Pass (official Austrian document on the health status of the child). Any kind of ophthalmological problem apart from blindness is coded as “visual impairment”; the need for glasses is coded as “visual impairment with visual aid.” In addition, at the age of 5 years, a neurodevelopmental assessment is carried out by a certified, experienced clinical psychologist. The neurodevelopmental assessment includes the Kaufman Assessment Battery for Children-II (KABC-II) and the Beery-Buktenica Developmental Test of Visual Motor Integration (Beery VMI).
Retinopathy of Prematurity
ROP stages were defined according to the International Committee for the Classification of Retinopathy of Prematurity (ICROP) . ROP exams were done by trained paediatric ophthalmologists at the MUV, and decision to treat was made according to the Early Treatment for Retinopathy of Prematurity Cooperative Group recommendations. At the MUV, the inclusion criterion for ROP screening is GA 31+6 at birth or less. Screening starts in the 5th week of life and is then performed weekly or fortnightly, depending on the patients’ age, occurrence, and progression of ROP.
The Beery-Buktenica Developmental Test of Visual Motor Integration
VMI performance in this study was assessed using the Beery VMI. The Beery VMI is a well-established test to assess the degree to which participants can integrate their visual and motor skills. The test is approved for age 2–99 in the full version. There is a visual perception subtest focussing on how visual information is perceived and a motor coordination subtest focussing on fine motor control. The Beery VMI has a high interrater reliability (0.75–0.88) . It has been used in research and clinical use for decades and is internationally validated. A standard score between 90 and 109 is considered average since it encompasses 50% of the population .
The Kaufman Assessment Battery for Children-II
Cognitive performance in this study was assessed using the KABC-II focussing on the simultaneous and the sequential processing scales . The original version of the KABC was developed by Nadeen L. Kaufman and Alan S. Kaufman in 1979. In 2004, a new version was published with the KABC-II. The KABC-II is applicable from age 3–18. A score between 85 and 115 (+/−1 standard deviation) is considered average .
The analyses were carried out using SPSS Statistics 26 (IBM) and R version 4.0.5 in R Studio. Categorical variables were expressed as frequencies and percentages, while continuous variables were expressed as mean and standard deviation. Differences between the groups ROP and No-ROP were calculated using the χ2 for categorical variables and Wilcoxon test for the continuous one. An ANOVA was used to detect differences in VMI and KABC scores between the No-ROP group and the three observed ROP stages. Finally, a MANOVA was conducted to identify the effect of ROP considering important demographic and medical characteristics (gender, birth weight, GA, severe intraventricular haemorrhage, chronic lung disease, i.e., the infant received any supplemental oxygen at any time on the date of week 36, late-onset sepsis, conventional ventilation). The basic patient characteristics such as gender, birth weight, and GA were included as fixed factors in the model, as was the primary parameter of ROP. Other important medical conditions were selected stepwise using a linear regression model. According to the correlation among the considered variables and the values of tolerance and variance inflation factor, there was no sign of collinearity for the finally selected items. Significance of conclusions was assumed below a p value of 0.05.
Thousand three hundred sixty-five patients were admitted to our neonatal intensive care unit in the considered timeline (January 2009 to December 2014); of these, 353 were included in the statistical analyses (online suppl. Fig. 1; for all online suppl. material, see www.karger.com/doi/10.1159/000529594).
Of the study patients, 216 had no ROP, while 137 infants had ROP (23 had stage 1, 74 had stage 2, and 40 had stage 3). Among ROP stage 3 patients, 93% had received laser therapy and/or cryotherapy. None of the children had a stage higher than stage 3.
Comparison between infants with and without ROP showed that infants with ROP had lower body weight (752 g ± 173 g vs. 996 g ± 239 g; p < 0.01) and GA (25+4 ± 11 d vs. 27+6 ± 13 d; p < 0.01) at time of birth. They had significantly more often severe illnesses such as late-onset bacterial sepsis/meningitis (42% vs. 26%; p < 0.01) or chronic lung disease (33% vs. 13%; p < 0.01), and they stayed significantly longer in hospital (97 ± 26 d vs. 68 ± 24 d; p < 0.01) (Table 1).
Further descriptive characteristics related to each ROP stage are presented in online supplementary Table 1. It was notable that infants with ROP stage 3 were significantly more often exposed to supplemental oxygen (100% vs. 88%; p < 0.01) and conventional ventilation (98% vs. 30%; p < 0.01) after initial resuscitation than those of the No-ROP group (Fig. 1).
Developmental outcome was significantly poorer in patients with ROP stage 3 compared to the No-ROP group with regard to all Beery VMI scores and in ROP stage 2 with regard to the Beery VMI standard score and the motor coordination subtest: Beery VMI standard score: stage 3 84 ± 19, stage 2 89 ± 14 versus no-ROP 99 ± 14; both p < 0.01; visual perception score: stage 3 84 ± 21 versus No-ROP 99 ± 17; p < 0.01; motor coordination score: stage 3 74 ± 22, stage 2 87 ± 18 versus No-ROP 93 ± 15; p < 0.01 and p = 0.03. Moreover, information regarding VMI values related to different stages of ROP is presented in Figure 2 and Table 2. When comparing only patients with stage 3 ROP who required laser therapy with those who did not, no differences were found in the total VMI score and the visual perception score (online suppl. Fig. 2); however, a difference was found in the motor coordination score (laser therapy 71.33 ± 19.97 vs. ROP no laser therapy 105.00 ± 18.08; p = 0.008).
Developmental outcome was also significantly poorer in ROP stage 3 compared to the No-ROP group concerning both KABC-II scores analysed and in ROP stage 2 concerning the simultaneous processing score (simultaneous processing: stage 3 88 ± 20, stage 2 94 ± 13 vs. No-ROP 100 ± 14; both p < 0.01; sequential processing: stage 3 84 ± 18 vs. No-ROP 94 ± 14; p = 0.02).
After correcting for the demographic and medical characteristics described above, children with ROP still scored 6.7 points (p < 0.01) lower in the Beery VMI than those from the No-ROP group. Poorer performance was also found in the motor coordination test (−4.9 points; p = 0.045) but not in the visual perception test. For more detailed information and the effect of the other variables of the model, see Table 3. After correcting for the demographic and medical characteristics described above, children with ROP no longer scored significantly poorer in the KABC-II test.
The current study shows that VMI was poorer in preschool children after having suffered ROP stage 2 or 3 compared to preterm infants with ROP stage 1 or without ROP. This study is the first to show the impact of ROP on VMI abilities at preschool age, even after adjustment for other major medical conditions.
Similar to our study, Molloy et al.  found that severe ROP (≥ stage 3) was associated with poorer outcomes on a wide range of neurodevelopmental measures including Beery VMI and Bayley MDI in extremely preterm adolescents. However, contrary to our results, this was not confirmed after adjusting for other important variables such as severe intraventricular haemorrhage or periventricular leukomalacia. The authors conclude that although ROP may be a risk factor for later development, many other clinical conditions may contribute to explain the neurodevelopmental outcome and that severe ROP per se is a pathological condition often associated with the smallest and sickest child. Inconsistencies in the results compared with our study could be related to the different age at follow-up, small sample size, different categorization of ROP, and different percentages of some important clinical conditions such as BPD or cystic PVL, which was never observed in our cohort. Further, Pétursdóttir et al.  reported differences in VMI scores between VLBW infants previously screened for ROP and a control group of term infants. However, they found no significant correlation between different degrees of ROP and Beery VMI scores and subtest scores in the VLBW group. Nevertheless, in this case too, the VMI scores were determined in a smaller group of young adults. They were divided into no ROP, untreated ROP, and treated ROP, where patients with treated ROP (n = 13) had lower VMI scores, similar to our study, but not significantly lower, probably due to insufficient power.
Our entire patient population including patients with and without ROP had a mean Beery VMI standard score in line with the literature showing in general poorer VMI performance for preterm infants compared to term infants [20, 21]. Interestingly, preterm infants without ROP in our study had almost normal Beery VMI values.
In contrast to the Beery VMI scores, we found no significant differences between ROP and No-ROP regarding the KABC-II scales in the linear regression model. This is in line with the study by Todd et al.  who also found no association between severe ROP/treatment for ROP and cognitive outcome at 3 years of age using a multivariate regression analysis to correct for other important variables. Molloy and colleagues found that severe ROP was associated with poorer MDI scores on the Bayley Scale of Infant Development, but after adjusting for other important medical conditions, the effect was also no longer significant . Other studies showed an impact of ROP on the Bayley MDI, even after correcting for other medical conditions [10, 23, 24]. However, in contrast to us, most of these studies included ROP stages 4 and 5, which were not observed in our collective. Only Sveinsdottir et al.  found a significant impact of ROP stages 1 and 2 on Bayley MDI.
In summary, this study found a linear decline associated with different degrees of ROP for all three components of the VMI test. However, even though stage 2 ROP values were lower than stage 1 ROP values, they were still within the norm, in contrast to stage 3 ROP values, which were on average below the norm. Considering also a progressively worsening clinical picture of the infant (stage 1–3), according to the conclusions of the study by Molloy et al.  and the fact that significant differences were found in the total score and in the motor score but not in the visual subscore, one could argue that this is a problem occurring at a higher cognitive level where this information is processed and therefore probably represents a problem in the integration of information at the cortical level rather than a visual problem per se. Our regression model also confirms the significant effect of ROP on VMI skills but not on the visual perception test, where other clinical variables were more representative in explaining the outcome under consideration. Finally, even considering the results of our regression, ROP per se is a pathological condition often associated with the smallest and sickest infants. It is therefore difficult to fully distinguish how much of the outcome can be explained by the general clinical condition of the patients from the isolated effect of ROP as it is difficult to determine how cortical pathways are affected in relation to this clinical condition.
Strengths and Limitations
The strength of this study is the relatively large number of patients that allows to reasonably analyse the effect of the unique stages of ROP on the considered outcome. Another strength is that data were validated for international standards following the Vermont Oxford Network definition guidelines. Moreover, our department of neonatology follows a very robust ROP screening programme. In contrast to its strength, a limitation of this study is its retrospective character, which implies that specific ophthalmological information has not been documented at the time of the follow-up examination. Visual acuity was not measured in follow-up as we could not report information on socioeconomic status. Further limitation might be the absence of ROP stage 4 and 5 in our collective.
This study shows that ROP plays a major role in impacting VMI abilities and that this was true even after adjustment for important demographic and medical characteristics. Since the correlation of VMI and handwriting skills of children has already been demonstrated [12, 25], further research is needed to determine the long-term consequences at school age and beyond for this fragile population of interest, particularly considering the fact that VMI difficulties are not necessarily related to cognitive abilities at preschool age. For the future, we believe that preterm infants with ROP should be comprehensively screened for VMI at preschool age, that a more detailed ophthalmological follow-up would be desirable to get better information on the whole spectrum of the visual sequelae of ROP, and lastly, that further research should clarify the impact of ROP on school performance and possible learning difficulties and subsequent behavioural problems.
The completion of this study could not have been possible without all medical doctors, nurses, and psychologists at the Division of Neonatology, Pediatric Intensive Care, and Neuropediatrics at the Medical University of Vienna for cooperating, supporting, and performing examinations.
Statement of Ethics
This study protocol was reviewed and approved by Ethics Committee of the Medical University of Vienna (MUV), approval number 1554/2020 and conducted in accordance with the World Medical Association Declaration of Helsinki. Given the retrospective nature of the study, informed consent was not required. Data were collected retrospectively for the period under consideration after approval by the Local Ethics Committee.
Conflict of Interest Statement
The authors declare no conflict of interest.
This study did not receive any kind of financial support.
Conceptualization and data analysis: Daniel Lukas Zimmermann and Vito Giordano; data collection: Daniel Lukas Zimmermann, Lukas Unterasinger, Hannah Schned, Liselotte Kirchner, and Renate Fuiko; original draft preparation: Daniel Lukas Zimmermann, Vito Giordano, and Manfred Weninger; writing: Daniel Lukas Zimmermann, Vito Giordano, Manfred Weninger, and Angelika Berger; review and editing: Daniel Lukas Zimmermann, Vito Giordano, Manfred Weninger, Angelika Berger, Katrin Klebermass-Schrehoff, Hannah Schned, Renate Fuiko, Liselotte Kirchner, and Monika Olischar; and project supervision: Vito Giordano and Manfred Weninger.
Data Availability Statement
All data generated or analysed during this study are included in this article [and/or] its online supplementary material files. Further enquiries can be directed to the corresponding author.