A total of 486 neonates met current ROP screening criteria. Of these 486 infants, 390 underwent ROP screening, had a known ROP outcome, and were included in the data analysis. The GA and BW of the neonates were 29.3±2.9 w and 1302.9±416 g, respectively. There were 41 (10.5%) SGA, 312 (80%) AGA and 37 (9.5%) LGA neonates. Characteristics of the study population were represented at Table 1. There were 169 neonates identified who would not have been screened for ROP if only the G-ROP criteria had been applied, and 84 (49.7%) of them were diagnosed with any stage of ROP that did not require treatment. The GA and BW of those 84 neonates that were missed by G-ROP model were 30.3±1.59 w and 1489±217 g, respectively. Sixty one of those 84 neonates would have been screened for ROP, if AAP and AAO ROP screening criteria were used. Sensitivity and specificity of G-ROP criteria for predicting any stage ROP for whole group and subgroups were given at Table 2.
Table 1
Characteristics of the study population
| All cases (n=390) | SGA (n=41) | AGA (n=312) | LGA (n=37) | p |
Gender, n (%) | | | | | |
Female | 194 (49.7) | 19 (46.3) | 163 (52.2) | 12 (32.4) | 0.067* |
Male | 196 (50.3) | 22 (53.7) | 149 (47.8) | 25 (67.6) | |
Gestational age, week | | | | | |
Mean±SD | 29.3±2.9 | 32 ± 2.6 | 29.2 ± 2.7 | 27.5 ± 2.9 | <0.001** |
Median (Min-max) | 30 (21-34) | 33 (24-34) | 30 (22-34) | 28 (21-32) | |
Birth weight, g | | | | | |
Mean±SD | 1302.9±416.9 | 1196.1 ± 297 | 1298.8 ± 407.7 | 1456.1 ± 553.6 | 0.045** |
Median (Min-max) | 1290 (390-2850) | 1240 (390-1640) | 1290 (480-2510) | 1510 (630-2850) | |
Birth type, n(%) | | | | | |
NVD | 91 (23.3) | 8 (19.5) | 71 (22.8) | 12 (32.4) | 0.349* |
C/S | 299 (76.7) | 33 (80.5) | 241 (77.2) | 25 (67.6) | |
*Chi-square test, **Kruskal Wallis test |
AGA: appropriate for gestational age, C/S: cesarean section, LGA: large for gestational age, NVD: normal vaginal delivery, SD: standard deviation, SGA: small for gestational age |
Table 2
Sensitivity and specificity of G-ROP criteria for predicting any stage ROP for whole group and subgroups
| ROP (+), n | ROP (-), n | Total |
All cases | | | |
G-ROP (+), n | 177 | 44 | 221 |
G-ROP (-), n | 84 | 85 | 169 |
Total | 261 | 129 | 390 |
Sensitivity 67.8% |
Specificity 65.9% |
SGA-neonates | | | |
G-ROP (+), n | 16 | 5 | 21 |
G-ROP (-), n | 8 | 12 | 20 |
Total | 24 | 17 | 41 |
Sensitivity 66.7% |
Specificity 70.6% |
AGA-neonates | | | |
G-ROP (+), n | 146 | 38 | 184 |
G-ROP (-), n | 64 | 64 | 128 |
Total | 210 | 102 | 312 |
Sensitivity 73.2% |
Specificity 87.7% |
LGA-neonates | | | |
G-ROP (+), n | 15 | 1 | 16 |
G-ROP (-), n | 12 | 9 | 21 |
Total | 27 | 10 | 37 |
Sensitivity 55.6% |
Specificity 90% |
AGA: appropriate for gestational age, G-ROP: Postnatal Growth and Retinopathy of Prematurity, LGA: large for gestational age, ROP: retinopathy of prematurity, SGA: small for gestational age |
The G-ROP model identified all of the treated neonates (n=26, 100%, 95% CI 87.1%-100%), and the GA and BW of the them were 24.8±1.8 w and 756±183 g, respectively. In all three groups (SGA, AGA, LGA), G-ROP model was 100% successful for determining severe ROP. Sensitivity and specificity of G-ROP criteria for predicting severe ROP for whole group and and subgroups were given at Table 3. All the treated ROP patients were caught by GA and/ or BW criteria, recieved intravitreal bevacizumab, and a total of 50 eyes were treated (both eyes of 24 patients and only one eye of two patients).
Table 3
Sensitivity and specificity of G-ROP criteria for predicting severe ROP for whole group and subgroups
| ROP treatment (+), n | ROP treatment (-), n | Total |
All cases | | | |
G-ROP (+), n | 26 | 195 | 221 |
G-ROP (-), n | 0 | 169 | 169 |
Total | 26 | 364 | 390 |
Sensitivity 100% |
Specificity 46.4% |
SGA-neonates | | | |
G-ROP (+), n | 1 | 20 | 21 |
G-ROP (-), n | 0 | 20 | 20 |
Total | 1 | 40 | 41 |
Sensitivity 100% |
Specificity 50% |
AGA-neonates | | | |
G-ROP (+), n | 21 | 163 | 184 |
G-ROP (-), n | 0 | 128 | 128 |
Total | 21 | 291 | 312 |
Sensitivity 100% |
Specificity 44% |
LGA-neonates | | | |
G-ROP (+), n | 4 | 12 | 16 |
G-ROP (-), n | 0 | 21 | 21 |
Total | 4 | 33 | 37 |
Sensitivity 100% |
Specificity 63.6% |
AGA: appropriate for gestational age, G-ROP: Postnatal Growth and Retinopathy of Prematurity, LGA: large for gestational age, ROP: retinopathy of prematurity, SGA: small for gestational age |
These 390 infants underwent a total of 2659 ROP examinations (6.8 examinations per infant), of which 662 (24.9%) could be reduced by applying G-ROP model without missing severe ROP. Forty one SGA-neonates underwent a total of 208 ROP examinations, of which 56 (26.9%); AGA-neonates underwent a total of 2166 ROP examinations, of which 519 (23.96%); and LGA-neonates underwent a total of 285 ROP examinations, of which 87 (30.5%) could be reduced by applying G-ROP model without missing severe ROP.
When perinatal morbidities were added individually to the G-ROP model, the sensitivity of the model was increased for any stage ROP. Among all cases, the sensitivity of the model was highest when BPD was added to the model (92.9%), followed by NEC, hsPDA, IVH, sepsis, and RDS, respectively (91.5%, 89.1%, 83.6%, 78.2%, 74%, respectively). Sensivity and specificity of G-ROP model for any ROP when perinatal morbidities added to the model were given at Table 4. Fourty seven neonates (12%) had none of those aforementioned perinatal morbidities, and none of those 47 neonates had severe ROP. When G-ROP model applied to those 47 neonates, 13 of them would have ROP examinations. When all of the six perinatal morbidities were added to the G-ROP model together, 53 ROP examinations of those 13 neonates could be reduced without missing severe ROP, and the reduction of ROP examinations could be increased from 24.9–26.9%.
Table 4
Sensivity and specificity of G-ROP model for any ROP when perinatal morbidities added to the model
| Sensivity % | (95% CI) | Spesifity % | (95% CI) |
All cases | 67.8 | 61.9-73.2 | 65.9 | 57.4-73.5 |
Sepsis | | | | |
+ | 78.2 | 71.9-83.5 | 59 | 46.5-70.5 |
- | 38.2 | 27.6-50.1 | 72.1 | 60.4-81.3 |
RDS | | | | |
+ | 74 | 67.7-79.5 | 66.7 | 54.4-77.1 |
- | 43.4 | 30.9-56.7 | 65.2 | 53.1-75.5 |
hsPDA | | | | |
+ | 89.1 | 79.1-94.6 | 36.4 | 15.2-64.6 |
- | 61.2 | 54.3-67.8 | 68.6 | 59.8-76.3 |
NEC | | | | |
+ | 91.5 | 80.1-96.7 | 50 | 23.4-74.6 |
- | 62.7 | 56.1-69 | 67.8 | 58.8-75.7 |
BPD | | | | |
+ | 92.9 | 87.1-96.2 | 47.1 | 26.2-69 |
- | 44 | 35.9-52.5 | 68.8 | 59.7-76.6 |
IVH | | | | |
+ | 83.6 | 72.9-90.6 | 76.9 | 49.7-91.8 |
- | 62.4 | 55.4-68.9 | 64.7 | 55.6-72.8 |
BPD: broncopulmonary dysplasia, CI: confidence interval, hsPDA: hemodynamically significant patent ductus arteriosus, IVH: intraventricular hemorrhage, NEC: necrotizing enterocolitis, RDS: respiratory distress syndrome |