Scholarlink

Objective <p>This study aimed to identify risk factors for central respiratory failure (CRF) in neuromyelitis optica spectrum disorder (NMOSD) and investigate predictors of successful mechanical ventilator liberation in affected patients.</p> Methods <p>This retrospective cohort analysis included 280 active patients with NMOSD admitted between 2020 and 2024. CRF was defined as respiratory failure caused by direct neurological injury to the brainstem or high cervical spinal cord. Among these patients, 28 developed CRF and were matched with 57 non-CRF controls (<i>n</i> = 85). Multivariable Cox proportional hazards regression was used to identify independent CRF risk factors, and a nomogram was generated. Predictors of successful weaning (liberation &gt; 48 h) were examined in the 28 patients with CRF using descriptive statistics and exploratory Fine–Gray competing risk analysis.</p> Results <p>The incidence of CRF in NMOSD was 10.0% (28/280). Multivariable Cox analysis identified two independent predictors of CRF: elevated platelet–lymphocyte ratio (PLR) [hazard ratio (HR) = 1.003, 95% confidence interval (CI) 1.001–1.005; <i>p</i> = 0.003] and medullary lesion on magnetic resonance imaging (MRI) (HR = 5.23, 95% CI 2.03–13.45; <i>p</i> = 0.001). The nomogram showed good discrimination (C-index = 0.779) and enabled effective risk stratification. Among patients with CRF, 75.0% (21/28) achieved successful ventilator liberation. Patients who achieved successful weaning had a significantly shorter time to glucocorticoid treatment and higher baseline PLR than those who failed (both <i>p</i> &lt; 0.05). In the multivariate competing risk model, higher hemoglobin subdistribution HR (sHR = 1.02) was a positive predictor, whereas higher body mass index (BMI) (sHR = 0.76), AQP4-IgG seropositivity (sHR = 0.39), and sepsis (sHR = 0.17) were independent negative predictors.</p> Conclusions <p>Elevated PLR and medullary lesions are independent risk factors for CRF in NMOSD, which can be predicted by the generated nomogram. Furthermore, successful ventilator liberation is achievable in most patients with CRF, and early glucocorticoid treatment alongside monitoring of prognostic factors such as BMI, AQP4-IgG status, and infection may guide clinical management.</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

Risk Stratification and Ventilator Liberation in Neuromyelitis Optica Spectrum Disorder-Related Central Respiratory Failure

  • 期刊论文

  • 发表日期：

    2026年3月12日

查看全文
  • Hua Huang,
  • Yang Liu,
  • Jiaxin Li,
  • Yawen Liang,
  • Qiuhui Huang,
  • Shengri Chen,
  • Miao Wang,
  • Zhijian Liang

摘要

Objective

This study aimed to identify risk factors for central respiratory failure (CRF) in neuromyelitis optica spectrum disorder (NMOSD) and investigate predictors of successful mechanical ventilator liberation in affected patients.

Methods

This retrospective cohort analysis included 280 active patients with NMOSD admitted between 2020 and 2024. CRF was defined as respiratory failure caused by direct neurological injury to the brainstem or high cervical spinal cord. Among these patients, 28 developed CRF and were matched with 57 non-CRF controls (n = 85). Multivariable Cox proportional hazards regression was used to identify independent CRF risk factors, and a nomogram was generated. Predictors of successful weaning (liberation > 48 h) were examined in the 28 patients with CRF using descriptive statistics and exploratory Fine–Gray competing risk analysis.

Results

The incidence of CRF in NMOSD was 10.0% (28/280). Multivariable Cox analysis identified two independent predictors of CRF: elevated platelet–lymphocyte ratio (PLR) [hazard ratio (HR) = 1.003, 95% confidence interval (CI) 1.001–1.005; p = 0.003] and medullary lesion on magnetic resonance imaging (MRI) (HR = 5.23, 95% CI 2.03–13.45; p = 0.001). The nomogram showed good discrimination (C-index = 0.779) and enabled effective risk stratification. Among patients with CRF, 75.0% (21/28) achieved successful ventilator liberation. Patients who achieved successful weaning had a significantly shorter time to glucocorticoid treatment and higher baseline PLR than those who failed (both p < 0.05). In the multivariate competing risk model, higher hemoglobin subdistribution HR (sHR = 1.02) was a positive predictor, whereas higher body mass index (BMI) (sHR = 0.76), AQP4-IgG seropositivity (sHR = 0.39), and sepsis (sHR = 0.17) were independent negative predictors.

Conclusions

Elevated PLR and medullary lesions are independent risk factors for CRF in NMOSD, which can be predicted by the generated nomogram. Furthermore, successful ventilator liberation is achievable in most patients with CRF, and early glucocorticoid treatment alongside monitoring of prognostic factors such as BMI, AQP4-IgG status, and infection may guide clinical management.