To explore the diagnostic cutoff value of the Silverman Anderson score(SA) predict to preterm infant with respiratory failure(RF).

From January 2013 to December 2013, 160 cases of the preterm infants with RF who needed for oxygen therapy were randomly selected from the Huaian Maternity and Child Healthcare Hospital.SA scores of all preterm infants at admission were respectively recorded, while extracting radial arterial for blood gas analysis. The diagnosis of RF were determined by comprehensive analysis. The diagnostic cutoff of the SA predicting preterm RF were confirmed by receiver operating characteristic (ROC) curve.

RF rate was 63.8%.SA score was 5.24±1.22. The area under the ROC curve was 0.91, area of the standard error was 0.023. 95% confidence interval area was 0.866-0.955.There was statistically significance in SA value for the diagnosis of preterm RF(P=0.000). The higher value of SA was, the greater the likelihood of preterm RF was. When the dangers of missed diagnosis rate is equal to that of misdiagnosis rate, the optimal SA diagnostic cutoff value was 5.5, the sensitivity was 67.6%, misdiagnosis rate was 3.4% and missed diagnosis rate was 37.1%.When sensitivity and misdiagnosis were rate dominant, the optimal SA diagnostic cutoff value was 4.5, the sensitivity was 93.1%, misdiagnosis rate was 31.0% and missed diagnosis rate was 14.9%.

The SA score can early predict preterm RF. It might help the clinicians to rapidly assess the severity extent of preterm infants with RF and give appropriate measures to the preterm infants at the bedside.Therefore, it might improve the quality of treatment of the preterm infants with RF.