Scholarlink

Background <p>Our aim was to develop a quantitative model for immediately estimating the risk of death and/or brain injury in late-onset sepsis (LOS) in preterm infants, based on objective and measurable data available at the time sepsis is first suspected (i.e., time of blood culture collection).</p> Methods <p>Retrospective study on neonates ≤36 weeks’ gestation with a positive blood and/or cerebrospinal fluid culture after 72 hours from birth.</p> Results <p>Among 3217 preterm live births, 94 cases were included (median gestational age 26.5 weeks’ IQR 25.0;28.0), of whom 26 (27.7%) had poor outcomes (17 death; 9 brain injuries). Infants with poor outcomes showed lower postnatal age (11.5 vs 12.5 days, <i>p</i> &lt; 0.001), lower mean blood pressure (30.5 vs 43 mmHg, <i>p</i> &lt; 0.001) and higher lactate levels (4.4 vs 1.5 mmol/l, <i>p</i> &lt; 0.001). Our multivariable model showed good discrimination and calibration (<i>c</i> statistic=0.8618, Hosmer-Lemeshow <i>p</i> = 0.8532), stratifying the population into 3 groups: low-risk (sensitivity 97%, specificity 52%), middle-risk, and high-risk (sensitivity 77%, specificity 80%).</p> Conclusion <p>This predictive model performs well as a practical and easy-to-use tool to help clinicians early identify the sickest neonates who may benefit from timely and aggressive support (e.g., central line, haemodynamic assessment) and close monitoring (e.g., 1:1 nursing assignment, frequent reassessments).</p> Impact <p><UnorderedList Mark="Bullet"> <ItemContent> <p>We lack data to early identify the severity of neonatal late-onset sepsis in preterm infants.</p> </ItemContent> <ItemContent> <p>Delay in treatment contributes to poor prognosis.</p> </ItemContent> <ItemContent> <p>We developed a model for early prediction of poor outcomes (mortality and brain injuries).</p> </ItemContent> <ItemContent> <p>The model utilizes immediately available and measurable data at the time sepsis is first suspected.</p> </ItemContent> <ItemContent> <p>This can help clinicians in tailoring management based on individual risks.</p> </ItemContent> </UnorderedList></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Outcome prediction for late-onset sepsis after premature birth

  • 期刊论文

  • 发表日期：

    2025年1月17日

查看全文
  • Francesca Miselli,
  • Riccardo Cuoghi Costantini,
  • Melissa Maugeri,
  • Elisa Deonette,
  • Sofia Mazzotti,
  • Luca Bedetti,
  • Licia Lugli,
  • Katia Rossi,
  • Maria Federica Roversi,
  • Alberto Berardi

摘要

Background

Our aim was to develop a quantitative model for immediately estimating the risk of death and/or brain injury in late-onset sepsis (LOS) in preterm infants, based on objective and measurable data available at the time sepsis is first suspected (i.e., time of blood culture collection).

Methods

Retrospective study on neonates ≤36 weeks’ gestation with a positive blood and/or cerebrospinal fluid culture after 72 hours from birth.

Results

Among 3217 preterm live births, 94 cases were included (median gestational age 26.5 weeks’ IQR 25.0;28.0), of whom 26 (27.7%) had poor outcomes (17 death; 9 brain injuries). Infants with poor outcomes showed lower postnatal age (11.5 vs 12.5 days, p < 0.001), lower mean blood pressure (30.5 vs 43 mmHg, p < 0.001) and higher lactate levels (4.4 vs 1.5 mmol/l, p < 0.001). Our multivariable model showed good discrimination and calibration (c statistic=0.8618, Hosmer-Lemeshow p = 0.8532), stratifying the population into 3 groups: low-risk (sensitivity 97%, specificity 52%), middle-risk, and high-risk (sensitivity 77%, specificity 80%).

Conclusion

This predictive model performs well as a practical and easy-to-use tool to help clinicians early identify the sickest neonates who may benefit from timely and aggressive support (e.g., central line, haemodynamic assessment) and close monitoring (e.g., 1:1 nursing assignment, frequent reassessments).

Impact

We lack data to early identify the severity of neonatal late-onset sepsis in preterm infants.

Delay in treatment contributes to poor prognosis.

We developed a model for early prediction of poor outcomes (mortality and brain injuries).

The model utilizes immediately available and measurable data at the time sepsis is first suspected.

This can help clinicians in tailoring management based on individual risks.