A new clinical respiratory distress score for surfactant therapy in preterm infants with respiratory distress

The guidelines for surfactant therapy are largely based on studies done in developed coun1tries wherein the facility infrastructure, patient profile, and clinical practices are different from low- and middle-income countries (LMICs). Though SRT is widely practiced in developing countries, there exists variability in clinical practice. Our objective was to identify the factors which would predict the need of surfactant administration and develop a “clinical respiratory distress (RD) score” for surfactant administration in preterm neonates with respiratory distress. A prospective observational study was conducted in 153 preterm infants (260/7 to 346/7 weeks gestation) with respiratory distress who were managed with CPAP and/or surfactant where indicated. Gestation < 32 weeks, no antenatal corticosteroid (ANS), hypothermia at admission, Apgar score < 3 at 1 minute, and Silverman score > 2 at 2 hours were found to be the significant factors in predicting surfactant requirement in multivariate regression analysis. A seven point scale was developed and categorized into two categories as < 4 and ≥ 4. The sensitivity, specificity, PPV, and NPV were 67%, 87%, 86%, and 68%, respectively, with a cutoff score ≥ 4. The positive likelihood ratio was 5.07 (95% CI 2.71–9.48), and negative likelihood ratio was 0.38 (95% CI 0.28–0.52). The observed rate of surfactant administration was found to be around 32% when the composite score was below four, and the rate increased to almost 86% when the composite score was ≥ 4. The predictive accuracy of the model was subsequently evaluated in a cohort of 56 preterm infants with respiratory distress.. Sensitivity, specificity and positive and negative predictive value during the validation phase were 97%, 73%, 85%, and 94%, respectively. With a composite score less than 4, the observed rate of surfactant administration was 6% (95% CI 1%–28%) as against the model predicted rate of 24%, while with composite score ≥ 4, the observed rate was 85% (95% CI 69%–94%) as against the model predicted rate of 90%.