Retinopathy of prematurity in very-low-birthweight neonates: epidemiology and risk factors

A retrospective study of 143 very-low-birthweight infants cared in a level III neonatal intensive care unit who had survived for at least 28 days. Initial eye ground evaluation was done at the postnatal age between 4 and 6 weeks. Follow-up evaluation was done every one to two weeks at the discretion of the ophthalmologists. Thirty-four variables were reviewed for each case. Statistical analysis was done for each variable, with the development of retinopathy of prematurity (ROP), severity of ROP and development of threshold ROP as the dependent variables, by Mann-Whitney U test or X2 test when adequate. Variables with P-valu < 0.05 were included in multiple regression. One hundred and thirty-eight cases were survived for more than 28 days with their eyes been checked. Twenty-six (18.8%) of them developed ROP. The prevalence of stage I was 2.2% (3/138), stage II was 3.6% (5/138), stage III was 12.3% (17/138), and stage V was 0.7% (1/138). Threshold disease, stage 3 (+) and above, was found in 5 cases (3.6%). Seventeen variables were found to be correlated with the development of ROP. Only the duration of continuous positive airway pressure (CPAP) was significantly correlated to the development of ROP in multivariate logistic regression. Fifteen variables were correlated with the severity of ROP, but only peak direct bilirublin level, peak total bilirubin level and duration of CPAP could entered multiple stepwise linear regression. Thirteen variables were correlated with the development of threshold ROP, but only episodes of septicemia enter the multivariate logistic regression. We postulate that the longer duration of CPAP in ROP cases may reflect the severity of apnea and episodes of hypoxic attacks. Reducing episodes of apnea may prevent the development of ROP. The number of episodes of septicemia was the only significant variable for threshold ROP so that infection control is important for the prevention of threshold disease.