Utility of Neonatal Sequential Organ Failure Assessment (nSOFA) Score for Neonatal Mortality Prediction

Abstract

Aims and Objectives:

(a) Utility of neonatal sequential organ failure assessment (SOFA) score for predicting mortality in neonatal sepsis. (b) Determining a cut-off score of neonatal sequential organ failure assessment (nSOFA) score in relation to mortality.

Materials and Methods:

Prospective pilot study in neonatal intensive care unit (ICU) for 1 year (January-December 2021).

Inclusion Criteria:

All neonates (inborn and outborn) admitted to neonatal ICU with the diagnosis of presumed or proven sepsis.

Exclusion Criteria:

Neonates with critical congenital heart disease, congenital malformations, congenital immunodeficiency syndromes, inborn errors of metabolism, and neonates with birth asphyxia. All neonates admitted in neonatal ICU with the diagnosis of presumed or proven sepsis were included in the study. Variables required for nSOFA were assessed along with general demographic like age, gestational age, and so on. The nSOFA score was documented twice, once at the time of admission (initial evaluation) for sepsis and later at 24 h. The final outcome (survivor/nonsurvivor) was recorded. Data was analyzed to see the correlation of nSOFA to outcome and also to see the predictive value of cut-off nSOFA score to outcome.

Results:

A total of 60 neonates with presumed or proven sepsis (30 each in term and preterm) were admitted, of which 10 (16.7%) were nonsurvivors (6 preterms and 4 term), P value .731. Four deaths were observed in outborn (n = 38) and 6 in inborn (n = 22) babies, P value .149. The mean (±standard deviation [SD]) of age (days) and weight (kg) among survivors was 11.24 (9.47) and 2.23 (0.69) and for nonsurvivors was 11.2 (11.9) and 1.9 (1.1), respectively, P value >.05. There was no statistical significance observed between survivors and nonsurvivors in relation to blood culture positivity, P value .17. Mean (±SD) nSOFA score among survivors at the time of admission (initial evaluation) was 1.96 (1.69) and at 24 h was 1.16 (1.48). Among nonsurvivors mean (±SD), admission (initial evaluation) nSOFA score was 7.6 (2.0) and at 24 h 10 (2) was observed, P value of <.001. The mean (±SD) of nSOFA changes (at the time of admission and at 24 h) for survivors (–)0.8(1.14) and for nonsurvivors 2.4 (1.17) was observed, P value <.001. Outcome in relation to nSOFA score, 90.0% mortality was observed with the score of ≥6 at the time of admission (initial evaluation) and score of ≥5 at 24 h. Based on these cut-off values, receiver operating characteristic curve for mortality prediction was calculated. Area under curve (AUC) for admission (initial) was 0.972 (sensitivity 90%, specificity 98%, positive predictive value 90%, negative predictive value 98, with accuracy of 96.7%). At 24 h, AUC was 0.999 (sensitivity 100%, specificity 98%, positive predictive value 90.9%, negative predictive value 100, with accuracy of 98.3%).

Conclusion:

nSOFA can predict neonatal mortality due to sepsis. Admission nSOFA score of ≥6 and ≥5 at 24 h correlate well with mortality.