Article Text

Download PDFPDF

Retinopathy of prematurity: screening and treatment in Costa Rica
  1. Ana Catalina Tabarez-Carvajal1,
  2. Milagro Montes-Cantillo1,
  3. Kelly H Unkrich2,
  4. Rupal H Trivedi2,
  5. Mae Millicent Winfrey Peterseim2
  1. 1 National Children’s Hospital of Dr Carlos Sáenz Herrera, San Jose, Costa Rica
  2. 2 Storm Eye Institute/Medical University of South Carolina Department of Ophthalmology, Charleston, South Carolina, USA
  1. Correspondence to Dr Mae Millicent Winfrey Peterseim, Department of Ophthalmology, Medical University of South Carolina, 167 Ashley ave, Charleston 29425-2503, South Carolina, USA; petersem{at}


Aim To determine the recent demographic data, risk factors and results of retinopathy of prematurity (ROP) screening and treatment in Costa Rica.

Methods The medical records of all preterm infants meeting ROP screening criteria (≤34 weeks’ gestational age (GA) or birth weight (BW) ≤1750g, and those determined at risk by neonatologists) in the national healthcare system, Costa Rica, January 2010–December 2014, were retrospectively reviewed. The numbers and percentages of infants with ROP, risk factors, percentage of patients treated and treatment outcomes were determined. Comparison is made with screening criteria and literature reports of ROP incidence in other countries.

Results The study population included 3018 preterm infants. Overall, 585 patients (585/3018, 19.4%) were found to have ROP. Of these, 15.4% (90 patients) required laser treatment, and 53% of those requiring treatment had BW <1000g. Five babies requiring treatment were ≥32 weeks’ GA but with BW ≤1750g. Aggressive posterior disease was found in nine patients, and two infants of those screened (2/3018, 0.07%) suffered severe visual impairment during the 5-year study period.

Conclusion We provide comprehensive data of ROP care in Costa Rica allowing assessment and comparison of screening criteria and protocol.

  • Retina
  • Child health (paediatrics)

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Advances in prenatal care allow survival of lower gestational age (GA) infants, leading to an epidemic of retinopathy of prematurity (ROP) in some low-income and middle-income countries.1 2 ROP screening and treatment programmes face challenges in low-income and middle-income countries due to difficult access and coordination of care.3 While reports from other Latin American countries exist, the incidence and outcomes of ROP in Costa Rica have not been reported.4

Under the leadership of Dr Mariana Vargas-Vargas and the National Children’s Hospital (NCH), Costa Rica began a nationwide ROP screening programme in 1982, with the development of a database in 2004. We provide the incidence and outcomes of the ROP programme based at the NCH of Costa Rica and compare screening criteria and results with those of other countries.


A retrospective chart review was carried out with approval of the Human Investigation Review Board of the NCH and the Bioethics Division of the Center for Strategic Development of Health Information and Social Security of the Costa Rican Department of Social Security or Caja Costarricense de Seguro Social, and was determined exempt from Human Investigation Review Board at the Medical University of South Carolina.

The national healthcare system, Caja Costarricense de Seguro Social, provides transportation for patients and families for medical visits. ROP screening was centralised to San Jose, with infants referred from 15 area hospitals and clinics. All infants ≤34 weeks’ GA or birth weight (BW) ≤1750 g, and those otherwise determined at risk by neonatology, are screened by NCH ophthalmologists every Friday morning. Treatment criteria and protocols adhere to the recommendations of the Early Treatment of Retinopathy of Prematurity (ETROP) Trial.5 A medical record for each preterm infant meeting ROP screening criteria and screened for ROP in Costa Rica is maintained at the NCH and was available to the investigators. Records from all infants born between January 2010 and December 2014 with GA ≤34 weeks and/or BW ≤1750 g and those referred directly by a neonatologist for ROP evaluation were reviewed. Infants were included if less than 37 weeks’ gestation and for whom complete and legible eye examination data were available.

All infants were evaluated in their first month of life, with subsequent visits as indicated by their disease course following ETROP guidelines.5 Infants were dilated prior to ocular examination with two drops of a 0.5% solution of tropicamide and 5% phenylephrine. Retina examination was carried out by a paediatric ophthalmologist trained at the NCH using a +28 dioptre lens. A lid speculum, scleral depressor and a drop of 0.5% tetracaine hydrochloride (Alcames) were used in some infants. Diode laser treatment was performed for eyes with type 1 ROP (zone I, any stage ROP with plus disease; zone I, stage 3 ROP with or without plus disease; or zone II, stage 2 or 3 ROP with plus disease) per ETROP recommendations within 72 hours by NCH ophthalmologists after administration of 0.5% tetracaine hydrochloride drops. Heart rate and oxygen saturation were monitored by neonatology personnel, with sedation administered on some infants as indicated.

Data relating to ROP screening and treatment, infant demographics, birth details and comorbidities were collected from the medical record and tabulated in an Excel table. Data included GA, BW, sex, hospital of origin, single or twin birth, most severe stage and grade of ROP, and any treatment of ROP. Infants were classified by the most severe stage of ROP recorded in either eye. Descriptive statistics methods were used. Percentages of infants screened and with ROP and the outcomes were determined. To facilitate the analysis and to be able to estimate the risk indicator (Odds Ratio), the factors of GA and BW were grouped into two categories. For GA, patients were divided into two groups: those ≤31 weeks’ GA and those >32 weeks’ GA. Infants were divided into two groups for BW: those ≤1500 g and those >1500 g.

Comparison was made with literature reports of ROP incidences in other countries. A subset of the data was then analysed for comparison with literature reports and to identify infants who would be missed if stricter criteria were used.


From January 2010 to December 2014, 3083 preterm infants were screened for ROP, and of these 3018 preterm infants met the inclusion criteria (<37 weeks with complete exam). A total of 585 patients (19.4%, 585/3018) screened were found to have ROP over the study period. Of these, 495 (84.6%, 495/585) resolved and 90 patients (15.4%, 90/585) required treatment (table 1). The annual percentage of infants screened and found to have ROP in Costa Rica declined from 18.1% (96/529) in 2010 to 17.4% (128/735) in 2014. The incidence of infants screened and found to have ROP requiring treatment decreased from 3.3% (18/529) in 2010 to 1.8% (11/628) in 2013, followed by an increase to 3.2% (24/735) in 2014 (table 1).

Table 1

Five-year survey of 3018 infants screened for ROP: no ROP, ROP that did not require laser treatment and ROP that required laser treatment

The mean GA of all infants was 30.7 weeks (SD: 2.4). For those without ROP, the mean GA was 31.2 weeks (SD: 2.2); for those with ROP and not treated, the average was 29.0 weeks (SD: 2.3), while those who received ROP treatment 28.0 weeks (SD: 2.4). The mean BW of all infants was 1394 g (SD: 353.2). The average BW of those without ROP was 1456.9 g (SD: 336.3), those with ROP and not treated was 1151.6 g (SD: 301.5 SD), and of those treated was 1035.6 g (SD: 260.4).

While most infants were referred from hospitals within the Central Valley, patients were sent to San Jose from all areas of the country (table 2, figure 1). The NCH, where the most fragile infants receive care, had a high percentage of preterm infants with ROP (58/198 infants, 29.3%).

Figure 1

Hospitals of the national health system, Caja Costarricense de Seguro Social in Costa Rica. Hospitals México, Niños, San Juan Dios, Calderón and Carit are within the San Jose Metropolitan area and offer expanded maternity services.

Table 2

Number of infants examined, with ROP, and treated by referring hospital within Costa Rica, 2010–2014

Infants were grouped into disease categories by the highest stage of ROP each infant attained and the zone (table 3). The most common disease was stage 1 in zone III (206/585 infants, 34.7%). The most common stage was stage 2 (265/585 infants, 45.3%). One patient suffered bilateral grade 5 ROP and one infant developed unilateral grade 4 disease.

Table 3

Number of infants per highest stage of ROP attained in each retina zone (percentage of total ROP cases)

Fifty-one per cent of those requiring treatment had BW <1000 g (46/90) (table 4). Forty three per cent of those requiring treatment had BW ≥1000 to ≤1500 g (39/90), and 5.6% had BW >1500 g (5/90).

Table 4

Number of infants and percentages by BW with ROP and laser treatment

Treatment-requiring ROP was noted in infants born as early as 23 weeks’ gestation and as late as 36 weeks’ gestation. The highest percentage of infants requiring laser treatment was found for infants born between 26 and 29 weeks of gestation (48.9%) (table 5).

Table 5

Number of infants and percentages by gestational age group screened with and without ROP and laser

Infants of ≤31 weeks’ GA were compared with those of >31 weeks’ GA for incidence of ROP. The risk (OR) was 4.946, with a 95% reliability index of 0.00. This indicates that preterm infants with GA ≤31 weeks have a fivefold higher risk of developing ROP than those of >31 weeks’ GA. Infants of BW ≤1500 g were compared with those of BW >1500 g. The risk (OR) was 5.407, with a 95% reliability index of 0.000. That is, premature babies with a BW ≤1500 g were 5.4 times more likely to develop ROP than those with BW >1500 g.

In order to be better able to compare with reports from other countries, the incidence of ROP was calculated, including only patients with BW ≤1500 g=26.6%, with GA <32 weeks=27.9%, and with GA ≤32 weeks and BW ≤1500 g=29.4%.


In 2010, an estimated 20 000 infants worldwide were blind from ROP, with 3500 progressing to severe visual impairment or blindness in Latin America.4 The rate of blindness from ROP in high-income countries is reported as <0.5%, while that in poor countries is greater than 1%–2%.6–8 The greater magnitude of blindness and visual impairment in low-income and middle-income countries is largely due to poor access to ophthalmic care and places a large burden on society.3 9–11 In the effort to eliminate avoidable blindness, the WHO and the International Agency for the Prevention of Blindness launched VISION 2020: The Right to Sight.12 13 This project includes measures to monitor and improve ROP screening and treatment programmes in low-income and middle-income countries.8–10 We provide an initial comprehensive report of ROP care in Costa Rica and compare our findings with recommendations and results in other countries.

ROP screening criteria in Costa Rica have changed over time.14–16 The VISION 2020 initiative for Latin American countries recommended screening infants with BW ≤1750 g and/or GA ≤32 weeks, and those deemed at higher risk by the neonatologist, with the recommendation to adjust locally in accordance with population, neonatal outcomes and levels of care for each neonatal intensive care unit.17 Screening criteria for ROP in Costa Rica were expanded to GA ≤34 weeks to ensure identification of all at-risk infants. It has been well documented that infants born in Latin American countries can develop ROP with increased GA and BW compared with the USA.3 4 9 13 Indeed, we found 9 babies at 31 weeks’ GA, 5 babies at 32 weeks’ GA and one at 33 weeks or more who required treatment (table 5). Four infants requiring treatment would have been missed if using the criteria recommended in the USA (≤30 weeks and/or <1500 g), and one if using the criteria recommended in the UK (≤31 weeks and/or <1500 g) or Latin America. These infants’ BW ranged from 1530 g to 1650 g, with GA 31–36 weeks. These findings are useful as Costa Rica doctors evaluate current screening criteria.

In low-income and middle-income countries, increased survival of premature infants due to improved neonatal care results in an increase in ROP. The neonatal mortality rate (per thousand births) in Costa Rica declined over the study period from 6.82 in 2010 to 6.16 in 2014.18 It is possible that the trend towards increasing numbers of children screened for ROP (table 1) correlates with improved neonatal practices that allow younger, and more fragile, infants to survive. In addition, our results confirm the influence of extreme prematurity, low GA and low BW on the incidence of ROP.1 2

During the study period, 19.4% of babies screened developed ROP. Although comparison with other countries is difficult given a range of screening criteria and reported populations, the incidence of ROP in infants screened is within a range of reported incidences in other Latin American countries (table 6).4 19–24 For comparison, including only infants of GA ≤32 weeks and BW ≤1500 g, the incidence in Costa Rica of 29.4% is higher than that reported in Argentina (26.2%) and Brazil (24.2%).

Table 6

Comparison of incidence of ROP by country, year of report and number of included infants19–24

In developed countries, the incidence of ROP has been estimated to be 21.8% of babies with GA <32 weeks.25 For comparison, within Costa Rica, including only infants of GA <32 weeks, the incidence is higher at 27.9%. Comparative findings invite evaluation for possible risk factors and protocol variations.

The percentages of infants requiring treatment reported within Latin America are varied.4 6 18–22 Of those screened in Costa Rica, we find that 15.4% of babies with ROP received treatment, a higher percentage than in reported series within Latin America (table 6) and within the range reported for some developed countries, including England and Wales.25–28 These comparative results provide opportunity to examine current practices. However, they are limited by variations in population, severity of disease, screening and treatment criteria, and neonatal care, which play a role in reported disease and outcomes.

Two infants of those screened (2/3018, 0.07%) suffered severe visual impairment in the 5-year study period. This percentage is below the estimated 3.0% of babies less than 32 weeks’ GA who suffer visual impairment due to ROP worldwide25 and is comparable with those reported for established treatment programmes.6 28 In summary, these results suggest the current national ROP screening parameters (infants GA ≤34 weeks and BW ≤1750 g, and at the discretion of neonatology) are appropriate for Costa Rica and effective by worldwide standards. Future analyses will continue to evaluate current practices to further maximise the effectiveness of ROP care in Costa Rica.



  • Contributors ACTC and MMWP had access to all of the data and take responsibility for the integrity of the data and analysis. Study concept and design: ACTC, MMC and MMWP. Acquisition, analysis and interpretation of data: all authors. Drafting of the manuscript: KHU and MMWP. Critical revision of the manuscript for important intellectual content: all authors. Statistical analysis: KHU and RHT.

  • Competing interests None declared.

  • Patient consent There is no identifiable patient data included in this report.

  • Ethics approval Scientific Ethics Committee of the National Children’s Hospital, Costa Rica, Comite Etico Cientifico Hospital Nacional de Ninos.

  • Provenance and peer review Not commissioned; externally peer reviewed.