Scholarlink

<p>Red cell alloimmunization in pregnancy poses significant risks for maternal and fetal health, particularly in resource-limited settings. Early identification of at-risk women is crucial for optimizing interventions and resource allocation.&#xa0;This retrospective cohort study was conducted at a tertiary care center, enrolling 2,084 pregnant women who underwent routine antenatal red cell antibody screening from January 2020 to December 2021. Predictors evaluated included maternal age, hemoglobin concentration, ABO and Rh status, parity, and red cell antibody screening results. A multivariable logistic regression model, incorporating penalized methods for robust estimation and variable selection, was developed to predict clinically significant alloimmunization. Model performance was comprehensively assessed using the area under the receiver operating characteristic curve (AUC), calibration measures (Hosmer-Lemeshow test, calibration slope, and intercept), and decision curve analysis. Internal validation was rigorously performed via bootstrapping with 1,000 samples. Clinically significant alloimmunization was observed in 68 women (3.3%) within the cohort. Key independent predictors identified were increasing maternal age (adjusted Odds Ratio 1.13 per year; 95% CI 1.04–1.22; <i>p</i> = 0.003), Rh-negative status (aOR 5.30; 95% CI 2.22–12.68; <i>p</i> &lt; 0.001), and a positive antibody screening result (aOR 102.7; 95% CI 36.7–287.6; <i>p</i> &lt; 0.001). The final model demonstrated excellent discrimination with an AUC of 0.87 (95% CI 0.82–0.91) and strong calibration (slope 1.01, intercept − 0.01, Hosmer-Lemeshow <i>p</i> = 0.48). At the optimal probability threshold of 0.058, the model achieved a sensitivity of 77.9%, specificity of 96.7%, a positive predictive value (PPV) of 44.2%, and a high negative predictive value (NPV) of 99.2%. Decision and clinical impact curves indicated substantial net clinical benefit across relevant thresholds. Subgroup analyses confirmed robust performance among Rh-negative women (AUC 0.88) and multigravida (AUC 0.89), though the PPV for primigravida was lower (22%). Our internally validated model, based on readily available clinical variables, effectively predicts the risk of red cell alloimmunization in pregnancy. Its application can support selective antenatal screening and targeted intervention, particularly where universal screening is not feasible. Further external validation is warranted to enhance generalizability and inform widespread clinical decision-making.</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 Risk Prediction Model for Red Cell Alloimmunization in Pregnancy: A TRIPOD-Compliant Single-Centre Retrospective Cohort Study

  • 期刊论文

  • 发表日期：

    2025年12月5日

查看全文
  • Anubhav Gupta,
  • Meenakshi Gothwal,
  • Garima Yadav,
  • Pratibha Singh

摘要

Red cell alloimmunization in pregnancy poses significant risks for maternal and fetal health, particularly in resource-limited settings. Early identification of at-risk women is crucial for optimizing interventions and resource allocation. This retrospective cohort study was conducted at a tertiary care center, enrolling 2,084 pregnant women who underwent routine antenatal red cell antibody screening from January 2020 to December 2021. Predictors evaluated included maternal age, hemoglobin concentration, ABO and Rh status, parity, and red cell antibody screening results. A multivariable logistic regression model, incorporating penalized methods for robust estimation and variable selection, was developed to predict clinically significant alloimmunization. Model performance was comprehensively assessed using the area under the receiver operating characteristic curve (AUC), calibration measures (Hosmer-Lemeshow test, calibration slope, and intercept), and decision curve analysis. Internal validation was rigorously performed via bootstrapping with 1,000 samples. Clinically significant alloimmunization was observed in 68 women (3.3%) within the cohort. Key independent predictors identified were increasing maternal age (adjusted Odds Ratio 1.13 per year; 95% CI 1.04–1.22; p = 0.003), Rh-negative status (aOR 5.30; 95% CI 2.22–12.68; p < 0.001), and a positive antibody screening result (aOR 102.7; 95% CI 36.7–287.6; p < 0.001). The final model demonstrated excellent discrimination with an AUC of 0.87 (95% CI 0.82–0.91) and strong calibration (slope 1.01, intercept − 0.01, Hosmer-Lemeshow p = 0.48). At the optimal probability threshold of 0.058, the model achieved a sensitivity of 77.9%, specificity of 96.7%, a positive predictive value (PPV) of 44.2%, and a high negative predictive value (NPV) of 99.2%. Decision and clinical impact curves indicated substantial net clinical benefit across relevant thresholds. Subgroup analyses confirmed robust performance among Rh-negative women (AUC 0.88) and multigravida (AUC 0.89), though the PPV for primigravida was lower (22%). Our internally validated model, based on readily available clinical variables, effectively predicts the risk of red cell alloimmunization in pregnancy. Its application can support selective antenatal screening and targeted intervention, particularly where universal screening is not feasible. Further external validation is warranted to enhance generalizability and inform widespread clinical decision-making.