At 3 years, the incidence rate, cumulative incidence, and prevalence of all amblyopia risk factors in premature children did not significantly differ between those who did and did not qualify for acute-phase ROP screening. After 24 months, the prevalence of amblyopia risk factors in premature children who were not at risk for ROP was between 22% and 25%, which is higher than the reported prevalence in children of that age (15% to 20%).6 In our population, the prevalence was between 7% to 22% for amblyogenic refractive errors and 3% to 13% for strabismus. Accordingly, it may be appropriate to screen premature children around 24 months of age for amblyopia risk factors, including those not eligible for acute-phase ROP screening.
In the Context of the Current Literature
This study investigated amblyopia risk factor frequency in premature children regardless of eligibility for ROP screening during the first 3 years of life. A comparison of strabismus and refractive error frequencies between studies is limited because few studies include premature children who do not qualify for ROP screening and there are varying threshold definitions for significant refractive errors.
Our study population included a high proportion of Hispanic and African-American children (Table 1). Compared to the Multi-Ethnic Pediatric Eye Disease Study (MEPEDS) data on Hispanic and African-American children, we found a higher prevalence of astigmatism and hyperopia at 36 months in our non-ROP screened group. Astigmatism prevalence was reported to be 5.1% in African-American children and 10.7% in Hispanic children aged 36 to 47 months.13 We found a higher 36-month prevalence of astigmatism in our non-ROP screened group: 16% when defined as greater than +2.00 diopters (D) and 24% when defined as greater than +1.50 D. At 36 to 47 months, hyperopia of greater than +4.00 D affects 4.3% of African-American and Hispanic children.14 We found a higher 36-month prevalence of hyperopia in our non-ROP screened group: 8% when defined as greater than +4.50 D and 12% when defined as greater than +4.00 D. Chen et al.15 also found that hyperopia and astigmatism were more prevalent in premature children (gestational age of less than 35 weeks or birth weight of less than 1,500 grams), which included children without ROP, when compared to age-matched children. Thus, screening premature children particularly for hyperopia and astigmatism may be appropriate.
Strabismus has been found to be independently associated with prematurity.16 We found a higher prevalence of strabismus in both our non-ROP and ROP screened groups compared to the Baltimore Pediatric Eye Disease Studies (BPEDS) (12% to 17% vs 1.9% to 2.9% in the BPEDS).17 Our 36-month strabismus prevalence was also much higher than that found in the MEPEDS at 36 to 47 months of 1.9% to 2.6%.18
Limitations of our study include the lack of a control group of full-term children. We compared our findings to the MEPEDS and BPEDS baseline data, but there could be significant differences between study populations and amblyopia risk factor definitions that make comparisons of prevalence data challenging. Our cohort did not have a high incidence of severe ROP, which may predispose a child to significantly higher risk for amblyopia risk factors. Another limitation is multiple comparisons between the two groups.
We may have underestimated the frequency of refractive risk factors in our study as defined by Donahue et al.6 Refractive risk factor targets differ slightly between the AAO Preferred Practice Pattern7 and the 2013 AAPOS revised targets.6 We defined amblyopia risk factors according to the age-specific AAO guidelines because they aligned better with our age groups and differentiate the anisometropia subtypes.
In our study, compliance with outpatient comprehensive eye examinations decreased by 60% over the study period (Table 2) despite this being a high-risk population in a relatively captive health care system. Accordingly, our frequencies of amblyopia risk factors may be falsely high because patients who were not perceived as having vision problems may have decided not to return for recommended follow-up. However, the excluded patients had significantly lower gestational age (P = .006) and birth weight (P < .001) compared to those who were included in our study, which may underestimate our cumulative incidence of amblyopia risk factors. Furthermore, the demographic information for patients who had three or more examinations was not significantly different from those who had less than three examinations, with the exception of a history of intrauterine toxic exposure (correlation coefficient: −0.18, P = .029).
The proportion of patients lost to follow-up could also influence the generalizability of our study. To further explore the implications of low retention, we conducted a subanalysis that included only those patients who were evaluated at either the 6-, 12-, 30-, or 36-month examination (n = 53). In this sub-analysis, the 3-year prevalence of any amblyopia risk factor (20%), strabismus (13%), and any refractive error (10%) was similar to our 3-year total prevalence (Table 4). Nevertheless, this decline in follow-up is an important observation, and future studies are indicated to decipher how to best detect vision problems in these patients. Incorporating early, repetitive, and effective screening strategies to detect amblyogenic refractive errors19–21 in the primary care setting and in high-risk premature continuity clinics could offer a more efficient strategy to consider.
Finally, our study population was largely Hispanic and of lower socioeconomic status, and thus our results may be more generalizable to these populations. Given that Hispanic children have a higher risk of amblyopia risk factors, including astigmatism,13 hyperopia,14 and myopia,22 our study may overestimate the frequency of amblyopia risk factors. However, the 3-year cumulative incidence of any amblyopia risk factor in Hispanic patients at 36 months (22% with 95% CI: 0.42 to 0.78) was not significantly different than that in the non-Hispanic population (22% with 95% CI: 0.22 to 0.58) (P = .94).
This study demonstrated that at 3 years, the cumulative incidence, incidence rate, and prevalence of all amblyopia risk factors in premature children did not statistically significantly differ between those who were eligible for ROP examinations compared to those who were not. Vision screening in these patients around 24 to 30 months that focuses on detection of refractive errors may be warranted. A prospective study with a full-term control group would be useful to assess the difference in incidence of all amblyopia risk factors and to further assess the need for ophthalmologic examinations in premature children.