A predictive model for prognosis in very low birth weight infants with late-onset sepsis.

Objectives: This study aims to develop a predictive model to assess the probability of poor prognosis in very low birth weight infants (VLBWI) with late-onset sepsis (LOS).

Methods: A total of 309 eligible VLBWI with LOS were included in the study. Logistic regression was used to determine prognostic factors for VLBWI with LOS. A nomogram incorporating these factors was created to predict the probability of poor prognosis. Poor prognosis includes death and survival with severe complications.

Results: In the developmental cohort, the incidence of poor prognosis was 59.5% (147/247). Forward stepwise logistic regression analysis showed that HCO3, albumin (ALB), ionized calcium (iCa), blood urea nitrogen (BUN), gestational age (GA), and birth weight (BW) were independent predictors of poor prognosis in VLBWI with LOS. The predictive model showed good discrimination and calibration. In the developmental cohort, the prediction model had a sensitivity of 83.7%, a specificity of 74.0%, and a C-index of 0.845 (95% confidence interval: 0.795-0.894).

Conclusion: Our study identified independent predictors of poor prognosis in VLBWI with LOS and used them to construct a predictive model. This model can help clinicians to identify high-risk groups with poor prognosis early and provide important clinical reference information. Impact: This article highlights the development of a predictive model to assess the probability of poor prognosis in very low birth weight infants with late-onset sepsis (LOS). The model constructed in this manuscript was the first model to predict the poor prognosis of VLBWI with LOS. We mean a poor prognosis that includes death and some severe complications that may lead to long-term disability. Clinicians can use the model's scoring results to assess a patient's condition and accurately identify the occurrence of poor prognosis.