Scholarlink

Background <p>Acute poisoning is a frequent cause of emergency department admission and may rapidly progress to multiple organ dysfunction syndrome (MODS); however, existing scoring systems were not specifically developed for early MODS prediction in acute poisoning. This study aimed to develop and internally validate a parsimonious clinical score for early MODS prediction at emergency admission.</p> Methods <p>We retrospectively analyzed 362 patients with acute poisoning admitted between January 1, 2023, and June 1, 2025, and grouped them according to MODS occurrence during hospitalization. MODS was defined using the maximum daily SOFA score during hospitalization. Least absolute shrinkage and selection operator (LASSO) regression followed by multivariable logistic regression was used to develop the predictive model, with internal bootstrap validation. A simplified points-based score (P-GNLS) was derived from the final model. Receiver operating characteristic (ROC) analysis and decision curve analysis (DCA) were used to assess its predictive performance and compare it with existing scoring systems.</p> Results <p>Among the 362 patients, 76 (20.99%) developed MODS. In-hospital mortality was significantly higher in the MODS group than in the non-MODS group (39.47% vs. 1.40%, <i>p</i> &lt; 0.001) and increased with the number of dysfunctional organs. The LASSO–logistic regression model identified pesticide poisoning, Glasgow Coma Scale (GCS) score, neutrophil count (NEUT), blood lactate (Lac), and serum creatinine (SCr) as variables independently associated with MODS. These variables were used to construct the P-GNLS score (0–38 points). The P-GNLS score showed good calibration (Hosmer–Lemeshow χ²=10.7551, <i>p</i> = 0.2160) and discrimination (AUC = 0.972), and demonstrated greater net benefit than the comparator scores across clinically relevant threshold probabilities in decision curve analysis.</p> Conclusions <p>Patients with MODS following acute poisoning had substantially higher in-hospital mortality. In this single-center retrospective cohort, the P-GNLS score showed good discrimination and calibration for early prediction of MODS at emergency admission. It may help identify higher-risk patients, but external validation is required before broader clinical application.</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

Development and internal validation of a parsimonious clinical score for multiple organ dysfunction syndrome following acute poisoning: a retrospective cohort study

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月6日

查看全文
  • Shuhao Ye,
  • Jiujiu Gui,
  • Pinxuan Zhou,
  • Yuheng Shi,
  • Hu Tang,
  • Donghan Zhang,
  • Zhengkun Bao,
  • Yahui Tang,
  • Zhongqiu Lu

摘要

Background

Acute poisoning is a frequent cause of emergency department admission and may rapidly progress to multiple organ dysfunction syndrome (MODS); however, existing scoring systems were not specifically developed for early MODS prediction in acute poisoning. This study aimed to develop and internally validate a parsimonious clinical score for early MODS prediction at emergency admission.

Methods

We retrospectively analyzed 362 patients with acute poisoning admitted between January 1, 2023, and June 1, 2025, and grouped them according to MODS occurrence during hospitalization. MODS was defined using the maximum daily SOFA score during hospitalization. Least absolute shrinkage and selection operator (LASSO) regression followed by multivariable logistic regression was used to develop the predictive model, with internal bootstrap validation. A simplified points-based score (P-GNLS) was derived from the final model. Receiver operating characteristic (ROC) analysis and decision curve analysis (DCA) were used to assess its predictive performance and compare it with existing scoring systems.

Results

Among the 362 patients, 76 (20.99%) developed MODS. In-hospital mortality was significantly higher in the MODS group than in the non-MODS group (39.47% vs. 1.40%, p < 0.001) and increased with the number of dysfunctional organs. The LASSO–logistic regression model identified pesticide poisoning, Glasgow Coma Scale (GCS) score, neutrophil count (NEUT), blood lactate (Lac), and serum creatinine (SCr) as variables independently associated with MODS. These variables were used to construct the P-GNLS score (0–38 points). The P-GNLS score showed good calibration (Hosmer–Lemeshow χ²=10.7551, p = 0.2160) and discrimination (AUC = 0.972), and demonstrated greater net benefit than the comparator scores across clinically relevant threshold probabilities in decision curve analysis.

Conclusions

Patients with MODS following acute poisoning had substantially higher in-hospital mortality. In this single-center retrospective cohort, the P-GNLS score showed good discrimination and calibration for early prediction of MODS at emergency admission. It may help identify higher-risk patients, but external validation is required before broader clinical application.