Journal of Mazandaran University of Medical Sciences (Jan 2009)

Predictive value of SNAP-PE, SNAP, CRIB indices for prediction of disease severity and determination of death in infants admitted to NICU

  • Vajyheh Ghaffari Saravi,
  • Soghra Khani,
  • Mehrnoush Kosarian,
  • Hossein Zaeri aqamshhady

Journal volume & issue
Vol. 19, no. 73
pp. 2 – 9

Abstract

Read online

(Received 25 February, 2009 ; Accepted 28 December, 2009)AbstractBackground and purpose: There is growing interest in employing screening procedures to identify critical infants at the beginning of hospitalization at NICU and adjusting death rate with severity of early disease. Most of studies were carried out on preterm infants born in the same center. The aim of this study was to compare the diagnostic power of SNAP-PE, SNAP, and CRIB indices in determining the severity of disease and predict the mortality of neonates referred from other centers to NICU of Bou-Ali Sina Hospital, Sari, and compare them regardless of age at birth.Materials and methods: All newborns admitted the NICU with whatever age and birth weight less than 1500 grams compared with infants weighing 1500 grams or more during the first 24 hours of hospitalization by disease severity were evaluated by questionnaires SNAP-PE, SNAP, CRIB from January 2005 to April 2007. Patients due to death or early discharge in less than 24 hours or with lethal congenital anomalies were excluded from the study. The median score obtained in discharged or expired patients compared by Mann-Whitney U Test and the diagnostic power and differentiation of the three tools were compared using ROC curves. Cut point (cut off) of each tool was calculated and sensitivity, specificity, NPV, PPV were obtained respectively.Results: Out of 200 studied neonates, 60 patients (30%) expired. Median SNAP-PE, SNAP, CRIB was higher in the died patients. Appropriate cut-off point based on the ROC curve for the CRIB was 6 and 10 for SNAP and was 12 for SNAP-PE. The area under the curve was 0.918 for the CRIB, 0.886 for SNAP, and 0.89 for SNAP-PE and all three tools were statistically significant (P<0.000). Sensitivity and specificity SNAP-PE, SNAP, CRIB for predicting neonatal death in infants weighing less than 1500 grams was lower than infants weighing 1500 grams and more.Conclusion: SNAP-PE, SNAP, and CRIB are suitable methods for predicting death and determine the severity of the disease in neonates. Therefore using these methods is recommended even for patients who referred to the NICU. J Mazand Univ Med Sci 2009; 19(73): 2-9 (Persian).

Keywords