Purpose
To validate the FIRST-ROP algorithm in a medium-risk cohort of infants undergoing screening for retinopathy of prematurity (ROP).
Design
Retrospective diagnostic comparison.
Methods
Infants who did not meet criteria for microprematurity or nanoprematurity (gestational age ≥ 27 weeks and birth weight ≥ 800 g) were included. Outcomes assessed included rates of ROP diagnosis and treatment-warranted ROP (TW-ROP), as well as postmenstrual age (PMA) at first diagnosis and initiation of treatment. Paired binary detection outcomes between traditional ROP screening strategies and the FIRST-ROP algorithm, which defers screening until 34 weeks PMA, were compared using the McNemar test.
Results
Among 8564 infants screened, 3641 (36%) met inclusion criteria, of whom 604 (17%) developed ROP. Median PMA at first ROP diagnosis was 35.1 weeks (interquartile range [IQR], 34.0-36.6 weeks; range, 31.3-51.4 weeks). Median PMA at stage 3 diagnosis was 35.6 weeks (IQR, 34.0-37.0 weeks; range, 32.3-47.6 weeks). Median PMA at first treatment was 40.0 weeks (IQR, 37.1-41.4 weeks; range, 34.3-44.4 weeks). Fifteen infants (2%) required treatment, of whom 10 (67%) had type 1 ROP. Initiating screening at 34 weeks PMA significantly reduced examinations compared with standard screening ( P <.001), corresponding to an estimated 17% fewer visits per patient. No cases of TW-ROP would have been missed under the FIRST-ROP strategy (95% CI, 99.3%-100%).
Conclusions
Over 3 decades, no infants in this medium-risk cohort required treatment before 34 weeks PMA. These findings validate the FIRST-ROP algorithm and support safe deferral of the initial ROP examination to 34 weeks PMA, reducing screening burden while maintaining sensitivity for TW-ROP.
INTRODUCTION
W ith the rising rates of prematurity and the improved neonatal survival, retinopathy of prematurity (ROP) has become one of the leading causes of reversible childhood blindness in the world. Traditional screening guidelines recommend early and frequent examinations, a strategy that has successfully minimized missed diagnoses and enabled early treatment. However, ROP screening has remained challenged by limited staffing, inadequate physician compensation, and significant burdens placed on families due to frequent clinic visits and travel, altogether placing a substantial toll on clinical services and ophthalmologists.
In recent years, there has been growing interest in refining screening strategies to better match the risk profile of individual infants. Several models have been developed to reduce unnecessary examinations. These include WINROP, CHOP ROP, G-ROP, TWO-ROP, and most recently the FIRST ROP. The FIRST-ROP algorithm by Altamirano et al, advances this effort by looking at a cohort of medium-risk preterm infants, defined as those do not fall into categories of microprematurity or nanoprematurity (≥27 weeks’ GA and ≥ 800 grams BW). Their findings showed no infants meeting type 1 treatment criteria based on the ETROP were diagnosed at their first or second examination. Their algorithm proposes that in medium-risk infants, later initiation of screening, rather than the traditional schedule, could decrease examinations while maintaining safety. There is a need to validate these findings in different cohorts that represent a variety of ethnic backgrounds, treating physicians, and neonatologists.
Our study aims to validate the FIRST-ROP algorithm in a cohort of medium-risk preterm infants cared for at the Jackson Memorial Hospital Neonatal Intensive Care Unit (NICU) over 3 decades. By evaluating whether delayed initiation of screening reliably identifies infants who develop clinically significant disease, we seek to assess the model’s potential to optimize resource utilization while ensuring patient safety.
METHODS
PARTICIPANT SELECTION
This was a single-center retrospective diagnostic comparison study including infants screened for ROP from February 1990 to October 2019 at Jackson Memorial Hospital neonatal intensive care unit by a single pediatric retina specialist (A.M.B). In contrast to the FIRST-ROP study by Altamirano, which evaluated only a contemporary cohort (post-2017), we included both contemporary and older cohorts (1990-2019). Our aim was to assess not only the current applicability of the algorithm but also the extent to which it could have been applied historically. This broader inclusion provides a more robust evaluation across different eras of ROP care and strengthens our understanding of the algorithm’s generalizability. However, to align more closely with the FIRST-ROP study and ensure methodological consistency, we also conducted a sub-analysis restricted to the most recent decade of our data (2009-2019). The study was conducted in adherence with the Declaration of Helsinki and the Health Insurance Portability and Accountability Act with a waiver of informed consent due to the retrospective nature of this work.
Infants that did not meet the criteria for microprematurity (24-26 weeks GA and/or BW 600-799 grams) or nanoprematurity (<24 weeks GA and/or BW < 600 grams), , hence those born at ≥ 27 weeks GA and with a BW ≥ 800 grams were included in the study. Both inborn infants and those transferred from outside institutions were included, provided they met the gestational age and birth weight criteria. Also, patients required a minimum follow-up of 34 weeks postmenstrual age (PMA) to be included in the study. Data collected included the following variables: (1) GA at birth, (2) BW, (3) PMA at first ROP diagnosis—defined as the initial examination at which stage ≥ 1 ROP was observed–, (4) ROP zone and stage per the International Classification of ROP criteria, , (5) presence of plus disease—defined as dilation and tortuosity of the posterior retinal vessels in at least 2 quadrants, (6) number of examinations performed before first ROP diagnosis, (7) PMA at first ROP treatment, and (8) number of inpatient ROP examinations. PMA was defined as a sum of GA plus chronological age. The interval between retinal examinations was determined by the attending ophthalmologist according to each patient’s clinical condition.
Cases of treatment-warranted ROP (TW-ROP) were identified according to the Early Treatment for Retinopathy of Prematurity (ETROP) definition criteria for type 1 ROP, which included: (1) zone I with plus disease, regardless of the ROP stage; (2) zone I, stage 3 without plus disease; (3) zone II, stage 2 or 3 with plus disease. The main outcomes were (1) rates of ROP and TW-ROP, (2) PMA at which ROP was diagnosed and treated (3) the sensitivity of detection when comparing screening initiation at 34 weeks PMA vs the traditional schedule (4) rate of TW-ROP at the first and second inpatient examination, (5) number of examinations performed before the first ROP diagnosis, and (6) number of total inpatient examinations performed. We excluded patients with (1) unavailable data regarding the first ROP diagnosis date or findings, (2) unavailable data regarding the GA or BW, (3) infants who met single or both criteria for micro and/or nanoprematurity, and (4) infants without at least 34 weeks of PMA follow-up.
STATISTICAL ANALYSIS
All statistical analyses were performed using Stata 18 (StataCorp LLC, College Station, TX). Descriptive statistics included medians with IQRs (IQRs) for continuous variables and frequencies with percentages for categorical variables. The McNemar test was applied to cf paired sensitivities between screening strategies and Clopper–Pearson exact method to calculate the 95% CIs.
RESULTS
A total of 8561 infants were screened for ROP between February 1990 and October 2019; of them, 3634 (42%) met inclusion criteria (born at ≥ 27 weeks GA and weighing ≥ 800 grams with a minimum follow-up of 34 weeks PMA). Seventy-three infants (2.0%) had follow-up limited to 34 weeks PMA, and only 5 infants (0.14%) initiated screening at 34 weeks PMA. Of the 3634 infants in the cohort, 195 (5.4%) were transferred from outside facilities; among these, 125 (64.1%) were transferred after 34 weeks PMA. The cohort’s mean GA and BW were 31.0 ± 3.1 (IQR, 29.0-32.7 weeks; range, 27.0-41.4) and 1452 ± 579 grams (IQR, 1080-1635 grams; range, 800-4905), respectively.
INFANTS WITH RETINOPATHY OF PREMATURITY
ROP was detected in 604 (17%) out of 3634 infants with a total of 2351 examinations. Among the 604 infants diagnosed with ROP, 6 (1.0%) had follow-up limited to 34 weeks PMA. Within the ROP cohort ( n = 604), 27 infants (4.5%) were transferred from outside institutions. Among these, 16 (59.3%) were transferred after 34 weeks PMA. The median PMA at first ROP diagnosis was 35.1 weeks (IQR, 34.0-36.6 weeks; range, 31.3-51.4 weeks). Among infants transferred from outside institutions, the median PMA at first ROP diagnosis was 34.3 weeks (IQR, 33.9-36.0 weeks; range, 32.0-41.4 weeks). Overall, 27 of the 195 transferred infants (13.9%) developed ROP, compared with 578 of 3439 infants (16.8%) born in-house. A total of 363 (60%) infants were diagnosed with stage 1, 194 (32%) with stage 2, 41 (7%) with stage 3, and 6 (1%) with stage ≥ 3 in their most affected eye. The median PMA at stage 1 diagnosis was 35.0 weeks (IQR, 34.0-36.4 weeks; range, 31.3-49.1 weeks). The median interval from birth to stage 1 diagnosis was 6.1 weeks (IQR, 4.7-7.6 weeks). The median PMA at stage 2 diagnosis was 35.3 weeks (IQR, 34.1-36.7 weeks; range, 32.5-51.4 weeks). The median interval from birth to stage 2 diagnosis was 7.0 weeks (IQR, 5.8-8.4 weeks). The median PMA at stage 3 diagnosis was 35.6 weeks (IQR, 34.0-37.0 weeks; range, 32.3-47.6 weeks). The median interval from birth to stage 3 diagnosis was 7.7 weeks (IQR, 6.4-9.6 weeks). 15 (3%) patients were treated in total, 14 out of them with laser, 3 with anti-vascular endothelial growth factor (VEGF) and 1 with surgery. The median PMA for first treatment was 40 weeks (IQR 37.1-41.4 weeks; range 34.3-44.4 weeks), and the median interval from birth to the first treatment was 11.4 weeks (IQR 9.3-13.1 weeks; range 5.2-16.9 weeks) Table 1 .
TABLE 1
Infants with ROP Requiring Treatment ( n = 15).
| Patient | GA (w) | BW (g) | PMA at First ROP Diagnosis (w) | Interval Between Birth and First ROP Diagnosis (w) |
Worst Stage
OD OS |
Worst
Zone OD OS |
Plus
OD OS |
Type 1 ROP
OD OS |
PMA at Treatment (w) | Interval Between Birth and First Treatment (w) |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 32 | 1745 | 42.3 | 10.3 | 3 | 2 | yes | yes | 44.4 | 12.4 |
| 3 | 2 | yes | yes | |||||||
| 2 | 30 | 1050 | 37.0 | 7.0 | 2 | 2 | no | no | 42.3 | 12.3 |
| 3 | 2 | no | no | |||||||
| 3 | 27 | 955 | 34.0 | 7.0 | 2 | 2 | yes | yes | 40.9 | 13.9 |
| 3 | 2 | yes | yes | |||||||
| 4 | 27 | 825 | 33.4 | 6.4 | 3 | 2 | no | no | 43.9 | 16.9 |
| 2 | 2 | no | no | |||||||
| 5 | 27 | 1065 | 37.9 | 9.9 | 5 | 2 | yes | yes | 38.7 | 11.7 |
| 3 | 2 | yes | yes | |||||||
| 6 | 28 | 1075 | 30.9 | 4.9 | 3 | 2 | yes | yes | 37.7 | 9.7 |
| 3 | 2 | yes | yes | |||||||
| 7 | 30 | 1060 | 34.3 | 4.3 | 3 | 3 | yes | yes | 37.1 | 7.1 |
| 2 | 3 | no | no | |||||||
| 8 | 27 | 850 | 33.9 | 6.9 | 3 | 1 | yes | yes | 34.3 | 7.3 |
| 3 | 1 | yes | yes | |||||||
| 9 | 28 | 860 | 35.9 | 7.9 | 3 | 2 | yes | yes | 44.1 | 16.1 |
| 3 | 2 | yes | yes | |||||||
| 10 | 27 | 1000 | 35.7 | 8.7 | 0 | N/A | no | no | 40.1 | 13.1 |
| 3 | 2 | yes | yes | |||||||
| 11 | 27 | 835 | 36.6 | 9.6 | 3 | 2 | yes | yes | 36.6 | 9.6 |
| 3 | 2 | yes | yes | |||||||
| 12 | 35 | 1500 | 38.9 | 3.9 | 2 | 2 | no | no | 40.0 | 5.0 |
| 2 | 2 | no | no | |||||||
| 13 | 27.4 | 800 | 40.9 | 13.4 | 0 | 2 | no | no | 36.7 | 9.3 |
| 2 | 2 | no | no | |||||||
| 14 | 30 | 1770 | 41.4 | 11.4 | 4 | 2 | yes | yes | 41.4 | 11.4 |
| 4 | 2 | yes | yes | |||||||
| 15 | 28.7 | 1080 | 39.9 | 11.3 | 4 | N/A | yes | no | 40.0 | 11.3 |
| 4 | N/A | yes | no |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
