PURPOSE
To estimate the cumulative incidence of strabismus among children diagnosed with retinopathy of prematurity (ROP).
DESIGN
Retrospective cohort study.
METHODS
Participants were children and adolescents aged ≤18 years with a diagnosis of ROP, with or without treatment exposure. We defined treated ROP as receipt of laser photocoagulation and/or intravitreal anti–vascular endothelial growth factor (anti-VEGF) injection, whereas untreated ROP had no record of either intervention. We estimated the cumulative incidence of strabismus at 1, 3, and 5 years after ROP diagnosis using Kaplan–Meier analysis. We used Cox proportional hazards models to calculate adjusted hazard ratios (aHR) and 95% confidence intervals (CI), adjusting for demographic variables and ocular comorbidities. The main outcome measure was development of strabismus following ROP diagnosis.
RESULTS
The cohort included 27,720 ROP patients: 973 were treated, and 26,747 were untreated. cumulative incidence of strabismus was consistently higher among treated patients. At 5 years, cumulative incidence was 34.78% in treated patients compared with 22.14% in untreated patients, with esodeviations as the most common subtype, occurring in 22.35% of treated patients and 11.87% of untreated patients. Treatment exposure was independently associated with an increased hazard of strabismus (aHR = 1.58; 95% CI = 1.29-1.95; P <.001). Amblyopia (aHR = 2.58; 95% CI = 2.25-2.96; P <.001) and nystagmus (aHR = 1.76; 95% CI = 1.48-2.10; P <.001) were also independently associated with increased hazard.
CONCLUSIONS
ROP confers an increased hazard of developing subsequent strabismus, with higher cumulative incidence among children who underwent treatment for ROP.
Retinopathy of prematurity (ROP) is one of the leading causes of preventable visual impairment in children worldwide. , Its prevalence varies by region and socioeconomic context, reflecting differences in neonatal care, oxygen regulation, and screening practices. , Although advances in neonatal intensive care have improved the survival of preterm infants, their success has (by design) expanded the population at risk for ROP-related complications. In the United States (U.S.), national data showed that the incidence of ROP among premature infants increased from 4.4% in 2003 to 8.1% in 2019, with the highest rates observed among infants of non-Hispanic White race (42.8%), followed by Hispanic (26.4%) and non-Hispanic Black (21.9%) infants. These trends underscore the persistent public health burden of ROP despite improvements in neonatal survival, and highlight ongoing demographic disparities in disease occurrence.
Beyond immediate retinal effects, ROP frequently results in long-term visual sequelae that extend into childhood, including refractive error, amblyopia, and strabismus. ,, Among these, strabismus is particularly common and has been reported in from 6% to more than 68% of ROP survivors. The pathogenesis of ocular misalignment in this population is multifactorial and may relate to asymmetric visual input, anisometropia, sensory deprivation, or structural alterations of the posterior pole and anterior segment. Despite its high prevalence among this subpopulation, most studies evaluating strabismus in this population are limited by small sample sizes, single-center design, or short follow-up duration, ,, leaving the cumulative incidence and temporal course of strabismus in ROP to be poorly characterized.
Given these gaps, our study aimed to evaluate the cumulative incidence of strabismus among patients with ROP and to identify factors associated with its development, including the impact of treatment exposure.
METHODS
STUDY DESIGN
We conducted a retrospective cohort study by gathering pediatric patient information from the TriNetX database’s (TriNetX LLC) “U.S. Collaborative” network, which comprises more than 70 health care organizations across the U.S. This dataset uses a federated data platform that harmonizes diagnostic codes to common standards. As part of this process, the TriNetX platform internally converts International Classification of Diseases, Ninth Revision ( ICD-9 ) codes to their ICD-10 equivalents to ensure consistency and comparability across participating institutions.
The Institutional Review Board at the University of Arkansas for Medical Sciences exempted this study because the database contains only de-identified patient information. The study adhered to the tenets of the Declaration of Helsinki (2013 version) and was compliant with the US Health Insurance Portability and Accountability Act of 1996 (HIPAA).
COHORT COMPOSITION AND ELIGIBILITY CRITERIA
Retinopathy of prematurity groups
We used ICD-9 and ICD-10 codes to identify patients with a history of ROP (ICD-10 H35.1x) and included patients aged 18 years or less. We required at least one ophthalmology encounter within the study period (January 1, 2004, to January 1, 2024).
The cohort was divided into 2 groups according to treatment status: (1) treated ROP: patients who received any ROP-related intervention, including laser photocoagulation or intravitreal anti–vascular endothelial growth factor (anti-VEGF) injection, identified using corresponding Current Procedural Terminology (CPT) or medication codes; and (2) untreated ROP: patients with ROP and no record of any treatment procedure.
All diagnostic and procedural codes used in cohort construction and outcome definitions are listed in the Supplemental Table.
INDEX EVENT AND FOLLOW-UP
We defined the index event as the first recorded ROP diagnosis that satisfied all of the aforementioned inclusion criteria. We analyzed follow-up durations of 1, 3, and 5 years after diagnosis separately. We excluded patients who lacked sufficient follow-up to reach a given timepoint from that timepoint-specific analysis, but we included them in shorter-term analyses.
CLINICAL ENDPOINTS AND STATISTICAL ANALYSIS
The primary outcome was the cumulative incidence of strabismus at 1, 3, and 5 years following ROP diagnosis. We assessed both the overall occurrence of strabismus and specific subtypes, including exodeviations, esodeviations, and vertical deviations. We identified strabismus using ICD-9 and ICD-10 codes (378.x and H50.x). We defined exodeviations using the following codes: 378.1, 378.23, 378.24, 378.42, H50.1, H50.33, H50.34, and H50.52; esodeviations using 378.0, 378.21, 378.22, 378.41, H50.0, H50.31, H50.32, and H50.51; and vertical deviations using 378.31, 378.32, 378.43, H50.2, and H50.53.
We estimated cumulative incidence for each outcome using Kaplan–Meier survival analysis within the TriNetX Analytics platform. We performed separate analyses for the treated and untreated subgroups. For the treated group, patients were required to have a recorded treatment procedure (laser photocoagulation or intravitreal anti-VEGF injection) after the ROP diagnosis; untreated patients had no record of any ROP-related procedure.
We additionally used a Cox proportional hazards model to analyze the independent predictors of strabismus development and to evaluate the effect of ROP treatment exposure on the development of strabismus. The following covariates were analyzed: gestational age, sex, race (White, Black, Asian, Native Hawaiian or Pacific Islander, and other), ethnicity (Hispanic/Latino or non-Hispanic/Latino), amblyopia prior to strabismus diagnosis (ICD-9 368.0; ICD-10 H53.0), nystagmus prior to strabismus diagnosis (ICD-9 379.50-379.56; ICD-10 H55.0), refractive or accommodation disorders (H52), optic nerve disorders (H46-H47), lens disorders (H25-H28), injury of the eye and orbit (ICD-10 S05), and glaucoma (H40-H42). We reported adjusted hazard ratios (aHRs) with 95% confidence intervals (CIs) for the 5-year outcome.
We used descriptive statistics to summarize baseline characteristics, including means and standard deviations for continuous variables and proportions for categorical variables. All analyses were performed using the TriNetX Analytics platform (TriNetX LLC; accessed December 2025). We defined statistical significance as a 2-sided P value of <.05. We generated figures using RStudio (RStudio Team, RStudio v4.5.2).
To protect patient privacy, TriNetX automatically suppresses and rounds cohort sizes when fewer than 10 patients are identified for a specific outcome, thereby minimizing the risk of individual re-identification within small groups. As such, we did not report outcomes with 10 or fewer patients within our analysis.
SENSITIVITY ANALYSIS
Because administrative databases may capture the first recorded diagnosis rather than the true neonatal diagnosis date, we performed a sensitivity analysis restricting the cohort to children aged ≤4 years at the index date. In this analysis, Kaplan–Meier cumulative incidence estimates and Cox proportional hazards models were used to evaluate the robustness of the primary findings.
RESULTS
CHARACTERISTICS OF THE RETINOPATHY OF PREMATURITY COHORT
Our study cohort included 973 children with treated ROP and 26,747 children with untreated ROP. In the treated group, 57.55% were male (560 patients), 48.09% were White (468 patients), and 20.54% were Black (200 patients), with 59.20% identified as non-Hispanic (576 patients). Among the treated group, 436 (44.81%) underwent laser photocoagulation alone, 453 (46.56%) received intravitreal anti-VEGF injections alone, and 43 (4.42%) received combined laser and injection therapy. In contrast, the untreated group comprised 51.58% male (13,796), 50.55% White (13,521), and 14.26% were Black (3,813) patients, with 67.58% non-Hispanic (18,076). Amblyopia and nystagmus were more common among our treated cohort, with amblyopia present in 19.12% of treated patients vs 14.01% of untreated patients, and nystagmus present in 8.02% of treated patients vs 4.06% of untreated patients. Glaucoma was an infrequent comorbidity, occurring in 3.70% of treated patients and 1.78% of untreated patients. Table 1 presents further details for the demographic and comorbidity characteristics for each group .
TABLE 1
Baseline Characteristics of Treated and Untreated Retinopathy of Prematurity (ROP) Cohorts
| Baseline Characteristics |
Treated ROP
(n = 973) |
Untreated ROP
(n = 26,747) |
|---|---|---|
| Sex, n (%) | ||
| Female | 413 (42.45) | 12,951 (48.42) |
| Male | 560 (57.55) | 13,796 (51.58) |
| Unknown | 0 (0.00) | 8 (0.03) |
| Race, n (%) | ||
| White | 468 (48.09) | 13,520 (50.55) |
| Black or African American | 211 (21.68) | 7,478 (27.96) |
| Asian | 34 (3.49) | 688 (2.57) |
| American Indian or Alaska Native | 10 (1.02) | 203 (0.76) |
| Native Hawaiian or Other Pacific Islander | 13 (1.33) | 96 (0.36) |
| Other race | 109 (11.20) | 1,931 (7.22) |
| Unknown race | 129 (13.25) | 2,830 (10.58) |
| Ethnicity, n (%) | ||
| Hispanic or Latino | 239 (24.56) | 5,357 (20.03) |
| Not Hispanic or Latino | 576 (59.20) | 16,254 (60.77) |
| Unknown ethnicity | 158 (16.24) | 3,897 (14.57) |
n = Number of patients.
CUMULATIVE INCIDENCE OF STRABISMUS IN RETINOPATHY OF PREMATURITY PATIENTS BY TREATMENT STATUS
Among untreated ROP patients, the cumulative incidence of any strabismus reached 10.57% (2,827 patients) at 1 year, 18.72% (5,007 patients) at 3 years, and 22.14% (5,922 patients) at 5 years. At 5 years, esodeviations accounted for most cases (3,175 patients), followed by exodeviations (2,021 patients) and vertical deviations (254 patients).
In the treated ROP group, the cumulative incidence of any strabismus was consistently higher than in untreated patients, occurring in 155 patients (15.97%) at 1 year, 278 patients (28.55%) at 3 years, and 338 patients (34.78%) at 5 years. At 5 years, esodeviations were the most common subtype (217 patients), followed by exodeviations (127 patients). Vertical deviation outcomes were suppressed (<10 cases) in accordance with the privacy policy. Table 2 summarizes the cumulative incidence of strabismus outcomes across follow-up intervals stratified by treatment status.
TABLE 2
Cumulative Incidence of Strabismus in ROP Patients
| 1 Year, % (n) | 3 Years, % (n) | 5 Years, % (n) | |
|---|---|---|---|
| Any Strabismus | |||
| Treated ROP | 15.97% (155) | 28.55% (278) | 34.78% (338) |
| Untreated ROP | 10.57% (2,827) | 18.72% (5,007) | 22.14% (5,922) |
| Esodeviations | |||
| Treated ROP | 10.44% (102) | 17.64% (171) | 22.35% (217) |
| Untreated ROP | 5.66% (1,513) | 9.98% (2,669) | 11.87% (3,175) |
| Exodeviations | |||
| Treated ROP | 4.42% (43) | 9.32% (91) | 13.03% (127) |
| Untreated ROP | 3.18% (1,268) | 4.90% (1,678) | 5.9% (2,021) |
| Vertical Deviations | |||
| Treated ROP | — (<10 cases) a | — (<10 cases) a | — (<10 cases) a |
| Untreated ROP | 0.30% (80) | 0.59% (158) | 0.95% (254) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
