الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 195 |  |  |
| | | from | 195 | | |
Abstract
Assessment of illness severity and mortality has become widely used in NICUs; they could identify high-risk neonates for a particular intervention and estimate the probability of mortality and morbidity of a particular infant. Illness severity measurements in NICUs have traditionally used birth weight and Apgar scores, but the relationship between mortality and these parameters has been insufficiently precise to use for quality assessment.
This literature illustrated that different scores for predicting mortality and morbidity had been implicated upon neonates. Among these scores we applied SNAP-II, n.SOFA and MSNS scores on critically ill preterm neonates and compared between them regarding their prognostic ability to predict mortality & morbidity among the studied neonates.
These scores had the advantages of being: easy for use, can be applied early in the course of hospitalization, able to reproducibly predict mortality & different morbidities, and useful for all neonatal categories whatever the diagnosis. The components of scores were clinical data which were already available and can be applied shortly after birth.
The aim of this study was to assess the prognostic ability of SNAP-II, MSNS and n.SOFA scores as predictors of mortality and morbidity among critically ill preterm newborn at Neonatal Intensive Care Unit of Menofia University Hospital.
This prospective cohort study was conducted on 200 critically ill preterm neonates admitted to the NICU of Menofia University Hospital during the period from October 2020 to March 2022. This study emphasized on critically ill preterm neonates which comprise the vast majority of neonatal deaths in our NICU. We excluded (full-term
Summary
- 148 -
neonates, neonates with lethal congenital or surgical anomalies, neonates who died or discharged from NICU during the first 4 days of life and those who had been referred to other hospitals during the course of management).
All neonates incorporated in this study were subjected to full history taking, thorough clinical examination & laboratory investigations. We extracted demographic and clinical data, including gestational age, birth weight, sex, mode of delivery, maternal & neonatal risk factors, APGAR scores at 1& 5 minutes, vital signs, capillary refill time, urine output, calculation of PaO2/FiO2 & SpO2/FiO2 ratios, interventions done (extensive phototherapy or exchange transfusion, type of respiratory support, surfactant administration and the need for inotropes), initial diagnosis and complications occurred to enrolled neonates. Laboratory investigation included (CBC with differential count, ABG, RBS, serum electrolytes, CRP, blood cultures, coagulation profile, liver & kidney function tests and transcranial ultrasound).
Data of the three scoring systems were collected & calculated as arithmetic sum of points assigned to each score variables. Scores were applied within the first 12 hours of admission and at the 4th day. We followed up the cases till their discharge or death.
SNAP-II included 6 variables (mean blood pressure, temperature, PaO2 / FiO2, PH, occurrence of seizures and urine output). Neonatal SOFA included the sum of [respiratory (intubation & SpO2/FiO2), cardiovascular (use of inotropes & systemic corticosteroids) and hematological scores (platelet count)]. MSNS included 8 variables (respiratory effort, heart rate, temperature, capillary refill time, random blood sugar, SpO2, gestational age and birth weight).
Summary
- 149 -
SNAP II and n.SOFA scores are directly proportional to the risk of mortality & morbidity i.e. the higher the score, the worse the prognosis, hence the higher the risk of mortality & morbidity. Whereas, MSNS score is inversely proportional to mortality & morbidity, therefore the lower the MSNS score, the higher the probability of mortality & morbidity.
These data were analyzed, tabulated and ROC curves were constructed to detect the prognostic performances of the three scores as predictors of neonatal mortality and morbidity. Relation and correlation between the scores and clinical data, therapeutic interventions and each diagnosis were done.
Results of this study revealed that 60.5% of neonates were males & 9.5% females, their mean gestational age was 32 weeks and mean birth weight 1685 grams. Most of them were high risk pregnancy with maternal and/or neonatal risk factors so, 75% were delivered by CS and 25% by normal vaginal delivery.
Concerning SNAP-II score and outcome:-
SNAP II score at admission had the highest predictive ability to detect mortality more than the other two scores. SNAP-II score at a cutoff point > 40 had (88.32%) sensitivity, (100%) specificity, (100%) positive predictive value, (86.5%) negative predictive value, (18.04) positive likelihood-ratio and (0.07) negative likelihood-ratio with the highest AUC value (0.968) indicating that it has the highest predictive ability to detect mortality among the other scores.
Both areas under the ROC curves for mortality & morbidity prediction at admission and at the 4th day were nearly similar. Also, there was no statistically significant difference between them (P value > 0.05).
Summary
- 150 -
Concerning n.SOFA score and outcome:-
Neonatal SOFA score at admission had the highest predictive ability to detect morbidity more than the other two scores because it had the highest AUC (0.956). The score at a cutoff point > 5 had (92.86%) sensitivity, (88.15%) specificity, (89.7%) positive predictive value, (94.2%) negative predictive value, (14.26) positive likelihood-ratio and (0.08) negative likelihood-ratio.
Both areas under the ROC curves for mortality & morbidity prediction at admission and at the 4th day were nearly similar. Also, there was no statistically significant difference between them (P value > 0.05).
Concerning MSNS score and outcome:-
It had the least predictive ability in detecting mortality and morbidity whether at admission & 4th day. It had the least AUC values among the other scores, they were (0.839, 0.836) at admission and at the 4th day respectively for morbidity ROC curve, and (0.845, 0.813) at admission and at the 4th day respectively for mortality ROC curve.
Regarding mortality; MSNS score at a cutoff point ≤10 had (73.68%) sensitivity, (58.21%) specificity, (67.8%) positive predictive value, (74.3%) negative predictive value, (6.55) positive likelihood-ratio and (0.17) negative likelihood-ratio. Whereas regarding morbidity; MSNS score at a cutoff point ≤10 had lower sensitivity (80.32%), specificity (76.7%), positive predictive value (66.4%), negative predictive value (72.1%), positive likelihood-ratio (8.67) and negative likelihood-ratio (0.26) than the other scores.
Both areas under the ROC curves for mortality & morbidity prediction at admission and at the 4th day were nearly similar. Also, there was no statistically significant difference between them (P value > 0.05).
Summary
- 151 -
Logistic regression analysis revealed that gestational age, birth weight and APGAR score at 5 minutes were important predictors of neonatal mortality, they were highly significantly associated with neonatal mortality (P <0.001). Meanwhile, intubation and the use of mechanical ventilation were found to be highly significant predictors of neonatal morbidity (P <0.001).
Our study concluded that SNAP-II and n.SOFA scores had excellent predictive ability to detect mortality which was slightly higher in SNAP-II score. MSNS score also had a good prognostic value regarding mortality prediction with lower sensitivity, specificity positive and negative predictive values than the other two scores.
Whereas, regarding morbidity prediction; n.SOFA was the best among the remaining scores with the highest sensitivity, specificity, positive and negative predictive values. SNAP-II and MSNS had good predictive ability to detect morbidity with lower sensitivity, specificity, positive and negative predictive values than n.SOFA.