Article Text

Download PDFPDF

Retinopathy of prematurity in Saudi Arabia: incidence, risk factors, and the applicability of current screening criteria
  1. A A Binkhathlan1,
  2. L A Almahmoud2,
  3. M J Saleh2,
  4. S Srungeri2
  1. 1
    King Fahad Medical City, Riyadh, Saudi Arabia
  2. 2
    Children Hospital, Riyadh Medical Complex, Saudi Arabia
  1. A A Binkhathlan, King Fahad Medical City, PO Box 21078, Riyadh 11475, Saudi Arabia; afaf{at}


Aims: To study the risk factors for Retinopathy of Prematurity (ROP) and the applicability of the current ROP screening criteria in Saudi Arabia.

Methods: A retrospective study of ROP incidence was conducted in patients of a neonatal intensive care unit in Riyadh from July 2003 until July 2004. Infants born at <36 weeks of gestation, and/or weighing <2000 g at birth, had their charts reviewed for ROP diagnosis and risk factors for ROP. The sensitivity and specificity of current screening criteria were assessed.

Results: One hundred and seventy-four infants were examined. Retinopathy of prematurity was diagnosed in 93 infants (56%); 15% of those patients were in stage 3 of the disease (severe ROP). The mean gestational age (GA) was 30 weeks for the ROP-positive group. At ⩽32 weeks’ gestational age and ⩽1500 g birth weight, the sensitivity of the current screening criteria was 68%, and the specificity was 55%. The most significant independent risk factor for the development of ROP was gestational age at birth.

Conclusions: This study found an older mean GA in infants developing ROP; it is recommended that the current screening criteria be widened to include 34-week GA infants into the programme. A tighter control on oxygen therapy is also recommended.

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.

In the last four decades, Retinopathy of Prematurity (ROP) has emerged as one of the leading causes of blindness in children in the developed world. The increased survival of extremely premature infants has led to a higher incidence of ROP. It is the second leading cause of blindness in the USA, accounting for 13% of students in blind schools.1 Recently, blindness due to ROP has become a major problem in developing countries, where the survival of premature infants has increased with limited O2 therapy monitoring. A recent report showed that the incidence of ROP related severe visual impairment or blindness in middle-income developing countries ranges between 0 and 38.6%.2

The Kingdom of Saudi Arabia (KSA) is considered a middle-income developing country, with a reported ROP incidence of 37.4%, and a threshold disease incidence of 9.7% in one centre.3 The current screening criteria, followed in KSA, identifies premature infants at a gestational age (GA) at birth of ⩽32 weeks, and a birth weight (BW) of ⩽1500 g, as at risk of ROP.4

Due to differences in neonatal care protocols and socio-economic status of each country, the ROP screening criteria followed in higher income countries might not be appropriate in a middle-income country, with a transitional economy, like KSA.5 In addition, infant health assessment varies in centres within the same country, leading to differences in ROP incidence rates in each centre. Recently, the International NO-ROP group suggested that each country should adopt their own screening programme based on the characteristics of infants with ROP seen in their own population.6


A retrospective, longitudinal cohort study was performed in a level III neonatal intensive care unit (NICU) in Riyadh, KSA. All admitted infants who survived until screening had their charts reviewed. The review period extended from July 2003 to July 2004. Infants were referred to the ophthalmologist running the screening clinic (primary investigator) for evaluation, and the CRYO-ROP group classification of ROP was followed.7 Follow-up exams were done every 2–3 weeks after the first evaluation until ROP regressed, the retina was fully vascularised without developing ROP, or ROP threshold disease was diagnosed, requiring laser therapy.8

Univariate logistic regression analysis was done to identify significant risk factors for ROP diagnosis. A stepwise logistic regression model was created, with backward elimination to determine the most important risk factor for the development of any stage of ROP. Risk factors that were included in the model were gestational age at birth, birth weight, amount of blood transfusion, number of days on mechanical ventilation and total number of days on oxygen therapy.

The sensitivity of the current screening criteria was calculated by dividing the number of infants with any stage of ROP as defined by the current criteria (infants ⩽32 weeks gestational age at birth, or birth weight ⩽1500 g combined), by the total number of infants with ROP in the study group, multiplied by 100. Similar calculations were performed for the proposed screening criteria of GA ⩽34 weeks at birth and ⩽1800 g birth weight.

The inclusion screening criteria followed in this study were <36 weeks of gestational age at birth and <2000 g of birth weight. Variables which were included in the statistical analysis were demographic data and risk factors, such as blood transfusions, number of days on mechanical ventilation, total number of days on O2 therapy, APGAR score at 5 min, intraventricular haemorrhage, Candida infection and sepsis. The first exam was performed at 32 weeks postmenstrual age (PMA = gestational age at birth+chronological age), or before discharge from NICU for more mature infants. Approval for the study was submitted to the Ministry of Health (MOH) ethical committee before collecting information from the infants’ files.


A total of 421 infants were admitted to the NICU during the study period; 174 were referred to the ophthalmologist to rule out ROP, and eight (5%) infants were removed from the final analysis due to missing files. The breakdown of infant admissions is shown in fig 1. There were 90 (54%) females and 76 (46%) males in the study group, and 124 (75%) of the infants were singletons, while 42 (25%) were the product of a multiple delivery. The mean GA at birth was 30.9 weeks (range 25–36 weeks), and the mean BW was 1411.5 g (range 650–3300 g).

Figure 1 Distribution of infant admissions to the NICU.

Retinopathy of prematurity was diagnosed in 56% (93/166) of the study group. Infants with ROP had a mean GA of 30.2 weeks (range 25–35) and a mean BW of 1325.7 g (range 650–3300). Two infants (1.2%) developed threshold ROP requiring laser therapy.

The distribution of infants according to GA and BW in relation to ROP diagnosis is shown in tables 1–3.

Table 1 Incidence of ROP, mean gestational age at birth, and mean birth weight
Table 2 Relationship of gestational age at birth to ROP diagnosis
Table 3 Relationship of birth weight to ROP diagnosis

Risk factor data were categorised for easier interpretation and used to identify independent risk factors for ROP in a univariate logistic regression analysis model. The significant risk factors for ROP diagnosis (p value of <0.05) were gestational age at birth (p<0.0001), birth weight (p = 0.0001), number of days under mechanical ventilation (p = 0.0026), total number of days on O2 therapy (p = 0.0021) and blood transfusion (p = 0.0017) (table 4).

Table 4 Independent risk factors for ROP

In multiple backward logistic regression analysis, all of the risk factors were included in the model. Gestational age at birth was found to be the most significant risk factor for ROP (odds ratio (OR) = 0.772, 95% CI from 0.652496 to 0.913577, p = 0.0026), followed by mechanical ventilation for >1 week (OR = 2.466, 95% CI from 1.115666 to 5.449754, p = 0.0257).

The sensitivity of the screening criteria, for the diagnosis with any stage of ROP, after combining GA and BW for the current criteria of GA ⩽32 weeks and BW ⩽1500 g, was 68%. The sensitivity for the suggested criteria of GA ⩽34 weeks and BW ⩽1800 g was 93%, while the specificity was 55% and 15% for both criteria respectively.

The current screening criteria when GA and BW were combined failed to identify 33 (35.5%) infants with ROP diagnosis, of which one was with stage 3 ROP.


The main objective of the present study was to evaluate the current screening criteria, to assess its applicability to general premature infant populations and to identify the general characteristics of infants at risk of visual disability. In this study, the use of current recommendations used for the screening of premature infants at risk for visual disability failed to detect all cases of ROP. In KSA, some larger infants developed blindness with a GA of 33 weeks and BW of 1800 g at birth (personal observations).

The current KSA screening criteria, when applied in this study, resulted in 30 infants with ROP being missed, including one severe ROP case (Stage 3). Similar findings were shown in Vietnam, India and China, with recommendations to include larger more mature infants in their screening programmes to avoid missing cases.911

In this study, ROP incidence within the KSA was higher than what was previously published by Dr Al-Amro, while the incidence of ROP threshold disease was lower.3 A comparison of this study with Dr Al-Amro’s is not possible, however, due to the differences in the general characteristics of infants studied, as well as the differences in the general medical care of the infants. The high rate of survival of extremely premature infants in Dr Al-Amro’s study likely explains the higher incidence of ROP threshold disease found in that study.

Guidelines for infant care in NICUs also need to be addressed, including tighter control and monitoring of supplemental oxygen to reduce the number of infants developing severe ROP.12 The risk factors found to be significant in this study population were similar to the reported risk factors in other populations.13 14 Inclusion of more mature infants of ⩽34 of gestational age at birth means examination at 4–6 weeks’ chronological age, with one exam being necessary to exclude the disease, which does not place a heavy load on the Ophthalmology clinic.15 Continuous evaluation of the screening criteria is necessary for each institution in the KSA to avoid missing infants developing visual disability.


This publication is one part of the dissertation for Master degree in the population health evidence from Manchester University. The primary and secondary investigators would like to extend their appreciations to Professor Dick Heller for his valuable comments, instructions, and support; he was a great teacher. We would like also to extend our many thanks to both reviewers from the BJO for their comments and suggestions, and the San Francisco edit for their editorial work.



  • Competing interests: None.