Prospective Evaluation of Serial Biomarkers in Patients With Intermediate High Risk Acute Pulmonary Embolism: A Single Center Proof‐of‐Concept Study

Figures

• 

  FIGURE 1. Study protocol timeline. Graphic depiction of the timing of sample collection during the study. Patients hospitalized for pulmonary embolism by CT angiogram meeting 2019 ESC criteria for IHRPE status were identified and consented for enrollment within the first 24 h of admission. Serial biomarkers and vital signs (blood pressure, heart rate, shock index, pulse oxygen saturation, and respiratory rate) were prospectively collected during seven time blocks over the first 3 hospitalization days. Time blocks from diagnosis were protocolized as follows: 0–8 h (time block 1), 8–16 h (time block 2), 16–24 h (time block 3), 24–36 h (time block 4), 36–48 h (time block 5), 48–60 h (time block 6), and 60–72 h (time block 7).

• 

  FIGURE 2. Biomarker concentration over time for Hemodynamic Decompensation. Median concentrations of TnT (A), NT‐proBNP (B), uric acid (C) and plasma lactate (D) were measured per protocol and patients divided by presence (N = 4, red line) or absence (N = 16, green line) of Hemodynamic Decompensation defined as persistent SBP < 90 mmHg or drop in SBP by 40 mmHg for more than 15 min with signs of end‐organ hypoperfusion (cold extremities, low urine output < 30 mL/h or mental confusion), use of vasopressors, or cardiopulmonary resuscitation. Shaded areas represent the interquartile ranges. The x‐axes represent prespecified time blocks, with tick marks indicating discrete sampling windows. The y‐axes display the measured biomarkers concentration in the original assay units. p‐values shown in the upper‐right corner within each panel reflect between‐group comparisons across time. Sample collection completeness differed by biomarkers and time blocks: for TnT and NT‐proBNP, time blocks 1 through 5 (0–8 h, 8–16 h, 16–24 h, 24–36 h, and 36–48 h), each included 20 patients while time blocks 6 and 7 each included 18 patients. For Uric acid, 0–24 h included 20 patients, 24–48 h included 20 patients,…

• 

  FIGURE 3. Biomarker concentration over time for Respiratory Decompensation. Median concentrations of TnT (A), NT‐proBNP (B), uric acid (C) and plasma lactate (D) were measured per protocol and patients divided by presence (N = 9, red line) or absence (N = 11, green line) of Respiratory Decompensation defined as the need for increase or persistent (≥ 3 days) respiratory support (higher than initial baseline supplemental oxygen, mechanical assisted ventilation) for SPO2 < 88% or increased work of breathing (RR ≥ 30 breaths per minutes with associated clinical signs of respiratory distress) or ratio of pulse oxygen saturation (SpO2) to a fraction of inspired oxygen concentration (FiO2) ≤ 260. Shaded areas represent the interquartile ranges. The x‐axes represent prespecified time blocks, with tick marks indicating discrete sampling windows. The y‐axes display the measured biomarkers concentration in the original assay units. p‐values shown in the upper‐right corner within each panel reflect between‐group comparisons across time. Sample collection completeness differed by biomarkers and time blocks: for TnT and NT‐proBNP, time blocks 1 through 5 (0–8 h, 8–16 h, 16–24 h, 24–36 h, and 36–48 h),…

• 

  FIGURE 4. Biomarker concentration over time for Clinical Deterioration. Median concentrations of TnT (A), NT‐proBNP (B), uric acid (C), and plasma lactate (D) were measured per protocol, and patients were divided by presence (N = 10, red line) or absence (N = 10, green line) of Clinical Deterioration, defined as hemodynamic or respiratory decompensation, need for advanced PE care, or death. Shaded areas represent the interquartile ranges. The x‐axes represent prespecified time blocks, with tick marks indicating discrete sampling windows. The y‐axes display the measured biomarkers concentration in the original assay units. p‐values shown in the upper‐right corner within each panel reflect between‐group comparisons across time. Sample collection completeness differed by biomarkers and time blocks: for TnT and NT‐proBNP, time blocks 1 through 5 (0–8 h, 8–16 h, 16–24 h, 24–36 h, and 36–48 h), each included 20 patients while time blocks 6 and 7 each included 18 patients. For Uric acid, 0–24 h included 20 patients, 24–48 h included 20 patients, and 48–72 h included 15 patients. For Lactate, 0–24 h included 20 patients, 24–48 h included 18 patients, and 48–72 h included 18 patients. See methods…

• 

  FIGURE 5. Biomarker concentration over time for Clinical Complication. Median concentrations of TnT (A), NT‐proBNP (B), uric acid (C) and plasma lactate (D) were measured per protocol and patients divided by presence (N = 16, red line) or absence (N = 4, green line) of Clinical Complication defined as prolonged hospitalization (longer than 4 days), major bleeding (fall in hemoglobin ≥ 4 gms/dL, red‐cell transfusion, or hemorrhagic stroke), hemodynamic or respiratory decompensation, need for advanced PE care, or death. Shaded areas represent the interquartile ranges. The x‐axes represent prespecified time blocks, with tick marks indicating discrete sampling windows. The y‐axes display the measured biomarkers concentration in the original assay units. p‐values shown in the upper‐right corner within each panel reflect between‐group comparisons across time. Sample collection completeness differed by biomarkers and time blocks: for TnT and NT‐proBNP, time blocks 1 through 5 (0–8 h, 8–16 h, 16–24 h, 24–36 h, and 36–48 h), each included 20 patients while time blocks 6 and 7 each included 18 patients. For Uric acid, 0–24 h included 20 patients, 24–48 h included 20 patients, and 48–72 h included 15…

• 

  FIGURE 6. Receiver‐Operator Characteristic curves for the outcome of Clinical Deterioration. A. Analysis of time from baseline to peak biomarker concentration to predict Clinical Deterioration. Clinical Deterioration defined as hemodynamic or respiratory decompensation, need for advanced PE care, or death. TnT AUC = 0.64 (95% CI = 0.42–0.82, p = NS). NT‐proBNP AUC = 0.82 (95% CI = 0.62–0.97, p < 0.01). The optimal cutoff determined by the Youden index was a time = 24 h, RR for outcome = 2.8×, specificity 90%, sensitivity 60%. Uric Acid AUC = 0.76 (95% CI = 0.49–0.97, p = NS). Plasma lactate AUC = 0.73 (95% CI = 0.45–0.94, p = NS). Blue solid line: performance of each biomarker for the outcome. Green dashed line: reference line. B. Analysis of Baseline PESI score to predict Clinical Deterioration. Clinical Deterioration is defined as hemodynamic or respiratory decompensation, need for advanced PE care, or death. PESI AUC was 0.755 (95% CI, 0.515–0.952), p = NS. Green dashed line: reference line.

• 

  FIGURE 7. Vital Sign Indicators over time for Clinical Deterioration. Serial vital signs were measured per protocol, and patients were divided by the presence (red) or absence (green) of Clinical Deterioration. The lowest SBP, SPO2, and highest HR in each time block were recorded. See Methods for additional details. Clinical Deterioration is defined as hemodynamic or respiratory decompensation, need for advanced PE care, or death. Shaded areas represent the interquartile ranges. Lines represent median values at each time point. Shaded areas represent the interquartile ranges (IQR). The x‐axes represent prespecified time blocks, with tick marks indicating discrete sampling windows. The y‐axes display the trended vital signs in the original units. p‐values shown in the upper‐right corner within each panel reflect between‐group comparisons across time. Sample collection completeness was similar for all vital sign parameters: time blocks 1 through 5 (0–8 h, 8–16 h, 16–24 h, 24–36 h, 36–48 h), each included 20 patients, while time blocks 6 and 7 included 18 patients. See methods and result sections for additional information. (A) Heart rate, (B) Systolic blood pressure, (C) Shock Index, (D)…