Pretreatment Haemodynamic Prediction of Recorded Early Circulatory Deterioration After Documented Short-Acting Beta-Blocker Treatment in Intensive Care: A Retrospective Multicohort Study
试验速览
- 阶段
- 不适用
- 状态
- 已完成
- 发起方
- 入组人数
- 2,541
- 试验地点
- 1
- 主要终点
- Percentage of ICCA Patients With Recorded Early Haemodynamic Deterioration
研究概览
简要总结
This retrospective observational study used existing intensive care records to examine whether heart rate and blood pressure measured before short-acting beta-blocker treatment could help predict early signs of worsening circulation recorded after treatment.
The study included patients who had received intravenous esmolol or landiolol during routine care. Researchers did not assign treatment or change patient care. The main analysis used records from Zhongshan Hospital, Fudan University. Additional intensive care databases were used to evaluate whether a prediction model developed at this hospital could be applied to other patient populations.
The main questions were whether pretreatment heart rate and blood pressure could predict recorded circulatory changes, whether combining these measurements improved prediction over simpler blood pressure measures, and whether differences in blood pressure monitoring affected the findings.
The main outcome combined a recorded fall in blood pressure, initiation or escalation of norepinephrine (a medicine used to support blood pressure), or early discontinuation of the same beta-blocker. Blood pressure and norepinephrine changes were assessed during the first three hours after the treatment index, and drug discontinuation during the first hour. External database analyses focused on recorded low blood pressure during the first three hours.
The study evaluated prediction among patients who had already received treatment. It was not designed to determine whether beta-blockers caused these events or whether a patient should receive a beta-blocker. Data collection and the analyses described here were completed before registration.
详细描述
Study rationale and design
This retrospective, multicohort observational study examined early recorded haemodynamic deterioration among intensive care patients with documented short-acting intravenous beta-blocker treatment. The purpose was to develop and evaluate a compact pretreatment prediction model using routinely collected physiological measurements. Treatment decisions were made during routine clinical care, independently of this study.
The development cohort was derived from the IntelliSpace Critical Care and Anesthesia (ICCA) system at Zhongshan Hospital, Fudan University. External evaluation used MIMIC-IV, the eICU Collaborative Research Database, AmsterdamUMCdb, HiRID, and the Salzburg Intensive Care Database (SICdb).
Analytical framework
The primary ICCA analysis used the first accepted, timestamp-verified esmolol or landiolol start per patient. Pretreatment predictors were median heart rate, diastolic blood pressure, and pulse pressure during the 120 minutes before the treatment index. Measurements recorded exactly at the index were excluded because their order relative to drug administration could not be established.
研究设计
- 研究类型
- 观察性
- 观察模型
- 队列研究
- 时间视角
- 回顾性
入排标准
- 性别
- All
- 接受健康志愿者
- 否
入选标准
- ICU patients with routinely collected medication and physiological records in ICCA at Zhongshan Hospital, Fudan University, or in MIMIC-IV, eICU, AmsterdamUMCdb, HiRID, or SICdb.
- Documented intravenous esmolol or landiolol treatment in ICCA, or documented intravenous esmolol treatment in the external validation cohorts.
- An identifiable treatment index according to database-specific medication-record rules. ICCA required a timestamp-verified treatment start. In eICU, the index was the first documented positive-rate infusion record rather than a verified initiation time.
- The last recorded mean arterial pressure within 30 minutes before the treatment index was at least 65 mmHg.
- At least one mean arterial pressure measurement was recorded after the treatment index and within 180 minutes.
- Documented ICU presence extended through 180 minutes after the treatment index.
- Pretreatment heart rate, diastolic blood pressure, and pulse pressure were available to calculate the model predictors within the 120 minutes before the treatment index.
排除标准
- Medication records that could not be mapped unambiguously to the patient, ICU stay, study drug, or treatment index.
- In ICCA, candidate starts with a conflicting same-time stop or an unresolved continuation or restart.
- No qualifying baseline mean arterial pressure measurement, or a last baseline mean arterial pressure below 65 mmHg.
- No mean arterial pressure measurement in the 180-minute post-index window.
- Insufficient documented ICU presence through 180 minutes after the treatment index.
- Missing pretreatment measurements required to calculate the model predictors.
- Record-selection rules:
- The first accepted treatment index per patient was selected before assessing outcome completeness in each cohort; later courses were not substituted when this first index failed eligibility requirements.
- Duplicate medication records at the same patient, stay, drug, and timestamp were collapsed.
- Physiological measurements recorded exactly at the treatment index were excluded because their temporal order relative to drug administration was uncertain.
研究组 & 干预措施
ICU Patients Receiving Short-Acting Intravenous Beta-Blockers
Patients receiving intravenous esmolol or landiolol during routine intensive care, identified retrospectively from six data sources. ICCA at Zhongshan Hospital, Fudan University, provided 1,354 esmolol- or landiolol-treated patients for model development and internal validation. External esmolol cohorts comprised MIMIC-IV (625), eICU (191), AmsterdamUMCdb (56), HiRID (88), and SICdb (227). The first eligible treatment index per patient was selected within each source. Eligibility required baseline MAP of at least 65 mmHg, complete pretreatment predictors, post-index MAP observation, and ICU presence through 180 minutes. ICCA assessed the recorded composite outcome; external cohorts evaluated the frozen MAP-only model without refitting, with AmsterdamUMCdb limited to descriptive evaluation. Treatment was not assigned by the study. Esmolol and landiolol were alternative recorded exposures, not a required combination.
干预措施: Esmolol (Drug)
ICU Patients Receiving Short-Acting Intravenous Beta-Blockers
Patients receiving intravenous esmolol or landiolol during routine intensive care, identified retrospectively from six data sources. ICCA at Zhongshan Hospital, Fudan University, provided 1,354 esmolol- or landiolol-treated patients for model development and internal validation. External esmolol cohorts comprised MIMIC-IV (625), eICU (191), AmsterdamUMCdb (56), HiRID (88), and SICdb (227). The first eligible treatment index per patient was selected within each source. Eligibility required baseline MAP of at least 65 mmHg, complete pretreatment predictors, post-index MAP observation, and ICU presence through 180 minutes. ICCA assessed the recorded composite outcome; external cohorts evaluated the frozen MAP-only model without refitting, with AmsterdamUMCdb limited to descriptive evaluation. Treatment was not assigned by the study. Esmolol and landiolol were alternative recorded exposures, not a required combination.
干预措施: Landiolol (Drug)
结局指标
主要结局
Percentage of ICCA Patients With Recorded Early Haemodynamic Deterioration
时间窗: From the documented treatment start to 180 minutes afterward; same-drug discontinuation assessed within the first 60 minutes.
Percentage of eligible ICCA patients with at least one recorded component of the composite outcome: (1) new mean arterial pressure (MAP) below 65 mmHg during (0,180\] minutes after the treatment index; (2) documented norepinephrine initiation, a within-source change from zero to a positive rate, or a greater than 20% increase from a positive pre-index rate during (0,180\] minutes; or (3) an explicit stop record for the same study beta-blocker during (0,60\] minutes. Norepinephrine changes were compared only within the same data source and measurement unit. Eligible patients had a last MAP of at least 65 mmHg during \[-30,0) minutes, post-index MAP observation, and ICU presence through 180 minutes. Patients meeting these observation requirements without a qualifying recorded signal were classified as having no recorded composite event. This outcome does not establish drug causality, and early stopping does not necessarily indicate intolerance. This composite was assessed in ICCA only.
次要结局
- Percentage of Patients With Newly Recorded Hypotension(Within 180 minutes after the database-specific treatment index)
研究者
研究点 (1)
标识符
- NCT 编号
- NCT07846995
- 其他研究编号
- B2026-484-BBlocker
日期
- 首次提交
- (7天前)
- 首次发布
- (昨天)
- 主要完成日期
- (10天前)
- 研究完成日期
- (8天前)
- 最近核实
- (29天前)
- 最近更新
- (昨天)
监管与共享
- 个体参与者数据共享计划
- 否
- 是否有结果
- 否