Scholarlink

Background <p>Variability in risk variable selection for prognostic scores/models after percutaneous coronary intervention (PCI) is poorly characterized. This systematic review examined how researchers report such variables and compared selection between Caucasian and non-Caucasian populations.</p> Methods <p>We searched PubMed/MEDLINE (inception to December 2025) for English-language studies developing, validating, or comparing risk scores/models for post-PCI prognosis. Data were synthesized descriptively. Risk of bias was assessed using PROBAST on a random sample.</p> Results <p>From 981 records, 212 articles (180 distinct scores/models) were included. Most frequent variables: sex/gender, age, diabetes, cerebrovascular disease, peripheral vascular disease, acute coronary syndrome, heart failure, Scr/eGFR, LVEF, and culprit artery. Significant differences between Caucasian (<i>n</i> = 94) and non-Caucasian (<i>n</i> = 86) scores/models were found for SBP, heart rate, Scr/eGFR, ACS presentation, ST-segment deviation, and culprit artery. PROBAST (20-model sample) showed high risk of bias in 70%.</p> Conclusions <p>Risk variable selection varies substantially across post-PCI prognostic scores/models, with significant ethnic differences. High risk of bias in existing studies highlights the need for more rigorous methodology.</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

A systematic review focusing on how medical researchers report variables in risk scores or models to predict prognosis of patients after percutaneous coronary intervention

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月11日

查看全文
  • Hong-Liang Zhao,
  • Jing Shi,
  • Guo-Qing Qi,
  • Ming-Qi Zheng,
  • Gang Wang

摘要

Background

Variability in risk variable selection for prognostic scores/models after percutaneous coronary intervention (PCI) is poorly characterized. This systematic review examined how researchers report such variables and compared selection between Caucasian and non-Caucasian populations.

Methods

We searched PubMed/MEDLINE (inception to December 2025) for English-language studies developing, validating, or comparing risk scores/models for post-PCI prognosis. Data were synthesized descriptively. Risk of bias was assessed using PROBAST on a random sample.

Results

From 981 records, 212 articles (180 distinct scores/models) were included. Most frequent variables: sex/gender, age, diabetes, cerebrovascular disease, peripheral vascular disease, acute coronary syndrome, heart failure, Scr/eGFR, LVEF, and culprit artery. Significant differences between Caucasian (n = 94) and non-Caucasian (n = 86) scores/models were found for SBP, heart rate, Scr/eGFR, ACS presentation, ST-segment deviation, and culprit artery. PROBAST (20-model sample) showed high risk of bias in 70%.

Conclusions

Risk variable selection varies substantially across post-PCI prognostic scores/models, with significant ethnic differences. High risk of bias in existing studies highlights the need for more rigorous methodology.