Holmstrom et al developed a risk equation for retinopathy of prematurity (ROP) using stepwise multiple logistic regression. They studied neonates in Stockholm, Sweden.
Population studied:
(1) 259 infants were in the initial group, with 202 single births and 57 from multiple births
(2) 33 infants were dropped out
Observations:
(1) The more immature the infant, the higher the risk of ROP.
(2) Gestational age at birth was more strongly associated with ROP than birthweight.
(3) In twins, sicker infants tended to have higher stages of ROP.
(4) About half the affected neonates had mild ROP (Stage 1 or 2) and half had severe ROP (mostly Stage 3).
(5) There was no correlation between gender and development of ROP.
Parameters found associated with ROP after the regression analysis:
(1) gestational age at birth in weeks
(2) presence of absence of bronchopulmonary dysplasia (BPD)
(3) birthweight in grams
Parameter
|
Finding
|
Points
|
bronchopulmonary dysplasia
|
absent
|
0
|
|
present
|
1
|
value of X =
= 10.26 – (0.304 * (gestational age in weeks)) + (2.303 * (points for bronchopulmonary dysplasia)) – (0.0019 * (birthweight in grams))
risk of retinopathy of prematurity =
= 1 / (1 + EXP((-1) * X))