A Personal Experience in Developing A Risk Prediction Algorithm

Selection of Patients for Systemic Thrombolysis in Pulmonary Embolism (PE)

Selecting the correct patient for systemic thrombolysis necessitates a thorough assessment of the patient’s preexisting comorbidities, mode of presentation and focused clinical examination to assess the immediate risk of hemodynamic collapse, the risk of long term complications, and the risk of major bleeding associated with the thrombolytic agent. As described above, high-risk PE patients warrant strong consideration of aggressive treatment options, including systemic thrombolysis with a high incidence of adverse outcomes if not instituted expediently. In patients who present with acute high-risk PE, the risk of mortality is high which makes the decision for systemic thrombolysis relatively easier as compared to the people who are hemodynamically stable. The case fatality of these hemodynamically unstable patients ranges from 35% to 58%. Therefore, benefits clearly outweigh the risk of adverse outcomes in the grand majority of high-risk PE patients who are not experiencing severe active bleeding.

Clinical Assessment:

Ideally, a prognostic model should be able to precisely identify the risk of mortality and recurrent PE in patients so that escalation of treatment can be performed when necessary. Also useful would be a risk model that predicted risks of various therapies beyond anticoagulation. Various risk predicting tools have been described in the literature with the Pulmonary Embolism Severity Index (PESI) (Table 1) being the best validated to determine short term mortality (30-day) in patients with PE. This prognostic model classifies patients from risk Class I (very low risk) to Class V (very high risk) based on demographics (age and sex), comorbidities (history of cancer, heart failure, chronic lung disease), and clinical findings (mentation, oxygenation, blood pressure, pulse and respiratory rate). Mortality risk ranged from 1% in Class I patients to 24.5% in Class V patients. With most patients falling in Class II and Class III, the negative predictive value reaches above 90% in low risk patients (Class I-III). Modified PESI was also introduced which is a simpler version of above described PESI with similar predictive precision but simpler to use. It includes age, history of heart failure or cancer and blood pressure, pulse rate and oxygen saturation.Both prognostic models can be used to risk stratify patients who can be eligible for thrombolytic therapy, but these models failed to predict the risk of adverse outcomes in these patients. In fact, there is no well-validated prediction model to assess the risk of bleeding in patients receiving thrombolysis for PE. Therefore, absolute and relative contraindications for thrombolytic therapy along with clinical judgment are the only tools available to risk stratify for bleeding.


Pulmonary Embolism severity Index (PESI) Simplified Pulmonary Embolism Severity Index (sPESI)
Demographics Demographics
Age >80 years Age >80 years
Male Sex
Comorbidities Comorbidities
History of Heart Failure History of cancer
History of cancer History of chronic lung disease
History of chronic lung disease
Clinical findings Clinical findings
Tachycardia>110 beats/min Heart Rate> 110/min
Systolic blood pressure <100 mm Hg Systolic blood pressure <100 mm Hg
Respiratory Rate >30/min Arterial oxygenation saturation <90% (with or without supplemental oxygenation)
Temperature <36 C
Altered mental status (lethargy, stupor, Coma)
Arterial oxygen saturation <90% (with or without supplemental oxygenation)

Table 1: Original and simplified pulmonary embolism severity index (PESI) (prognostic model to predict 30-day outcomes in patients with acute pulmonary embolism)


Biomarkers and Imaging assessment:

Rise in cardiac biomarkers, including troponin and brain-type natriuretic peptide (BNP) may represent right heart dysfunction and have been associated with an increased risk of PE related deaths. Chest CT scan is the gold standard imaging modality for patients who come to the ED with suspicion of PE. Easy-to-measure dimensions can be prognostically important in selecting patients who are at elevated risk of early deterioration.

Bedside transthoracic echocardiography can be utilized to detect RV dysfunction in the setting of acute PE. It can detect a  wide range of imaging indicators form very nonspecific right ventricular dilatation and hypokinesis to the very specific McConnell’s sign, in which the right ventricle has a characteristic appearance of significant enlargement with free wall dysfunction and relative apical sparing. Echocardiography should be performed to further risk stratify patients with clinical evidence of RV failure, elevated cardiac enzymes or in clinical decompensation. All these modalities can be helpful in not only classifying these patients as intermediate or high risk patients but also to segregate patients with high likelihood of early deterioration so that systemic thrombolysis can be utilized.

Summarizing the patient selection criteria for systemic thrombolysis, it is clear that the clinical judgment to utilize a thrombolytic agent in addition to anticoagulation in a patient with acute PE necessitates an individualized assessment of the benefits  of improving morbidity and mortality versus risk of major bleeding.


Absolute Contraindications Relative contraindications
§  Prior intracranial hemorrhage §  History of chronic severe uncontrolled hypertension
§  Known structural cerebral lesion §  Severe uncontrolled hypertension (systolic >180 mmHg or diastolic >110mmHg
§  Known malignant intracranial neoplasm §  History of ischemic CVA >3 months
§  Ischemic stroke within 3 months §  Trauma or prolonged CPR >10mins
§  Suspected aortic dissection §  Major surgery within 3 weeks
§  Active bleeding (excluding menses) §  Recent internal bleed (2-4 weeks)
§  Significant closed-head or face trauma within 3 months §  Pregnancy, active peptic ulcer, pericarditis, age>75years, diabetic retinopathy, recent invasive procedure, current anticoagulant use

Up to two third of acute PE patients do not receive thrombolytic therapy due to contraindications.

Table 2: Contraindications to systemic thrombolysis in acute pulmonary embolism patients


Often patients with acute pulmonary embolism (PE) do not receive thrombolytic therapy even if they could potentially benefit from such, mainly due to concerns for bleeding, especially intracranial hemorrhage (ICH).There are no known risk predictor algorithm(s) for assessing ICH risk in PE patients. A simple, novel risk score was developed to predict ICH in patients with PE treated with thrombolytics – the risk predictors including peripheral vascular disease (P), age>65 (Elderly-E), prior stroke (CVA-C) and prior heart attack (H)-components of the PE-CH score. The strongest risk predictor was a history of prior stroke, including both ischemic and hemorrhagic.

This is the first risk assessment algorithm to look at ICH risk in PE patients-future research should be directed towards refining this initial risk score by incorporating laboratory and radiographic parameters for optimal risk prediction. For patients, and healthcare providers alike, the risk score provides an easy opportunity to evaluate the risk of ICH for patients with acute PE being considered for thrombolytics. The PE-CH score provides an initial risk prediction model for ICH risk in patients with PE receiving thrombolytics. In patients with more than one risk predictor, especially those with a prior history of stroke, great caution must be exercised prior to consideration of thrombolytic use.



Thromb Haemost 2017; 117(02): 246-251
DOI: 10.1160/TH16-07-0588



When The Guidelines Need Guidance

I recently had the opportunity to be part of a team looking at the ‘evidence base and quality’ of recommendations enumerated in the current American Heart Association/American College of Cardiology guidelines for peripheral vascular interventions. The study led by my friend and colleague, Dr Partha Sardar, and Dr Herbert Aronow of the Warren Alpert Medical School at Brown University, found that the strength of evidence for the different recommendations vary significantly, underscoring the need for higher-quality evidence in this area as published in Circulation: Cardiovascular Interventions.

Our team identified 134 recommendations from five current full guidelines for endovascular and surgical procedures for peripheral vascular disease. For all peripheral vascular interventions, only 13% of recommendations were supported by level A evidence, whereas 48% were supported by level B evidence and 39% were supported by level C evidence.

The majority of recommendations were supported by level C evidence for pulmonary embolism or deep vein thrombosis interventions (76%) and inferior vena cava filter placement (69%), and level B evidence for renal artery stenosis interventions (67%).

However, levels of evidence were higher for endovascular therapy for stroke (level A, 24%; level B, 52%; level C, 24%), carotid revascularization (level A, 23%; level B, 52%; level C, 24%) and endovascular or surgical treatment for abdominal aortic aneurysm and lower-extremity aneurysm (level A, 26%; level B, 67%; level C, 7%). Quality of evidence for surgical revascularization for lower-extremity peripheral artery disease (level A, 18%; level B, 37%; level C, 45%) was also lower than for endovascular therapy (level A, 18%; level B, 55%; level C, 27%), which likely leads to greater emphasis on endovascular therapy in the current Appropriate Use Criteria (AUC for PAD) published by the national societies. (Bailey SR, Beckman JA, Dao TD, et al. ACC/AHA/SCAI/SIR/SVM 2018 appropriate use criteria for peripheral artery intervention: a report of the American College of Cardiology Appropriate Use Criteria Task Force, American Heart Association, Society for Cardiovascular Angiography and Interventions, Society of Interventional Radiology, and Society for Vascular Medicine. J Am Coll Cardiol. 2018;Epub ahead of print.)

Of significant surprise was the degree of variation in level of evidence supporting different procedural guideline recommendations. There was no level A evidence to support pulmonary embolism/deep vein thrombosis, inferior vena cava filter or renal artery stenosis intervention. In contrast, nearly 1 in 4 endovascular stroke therapy recommendations were supported by level A evidence.


Strength of recommendations

The researchers also noted that, overall, most recommendations were class II (54%), followed by class I (35%) and class III (11%).

For lower-extremity PAD endovascular revascularization, IVC placement, carotid revascularization and endovascular therapy for stroke, most recommendations were class II rather than class I or class III. For renal artery stenosis revascularization, recommendations were split evenly between class I and class II, with none falling into class III. For surgical or endovascular treatment of PE, there were no class I recommendations and 80% were class II. The classes of recommendation also varied for other peripheral vascular interventions, including DVT interventions, endovascular or surgical treatment for mesenteric artery disease, interventions for subclavian and brachiocephalic arteries, and endovascular or surgical treatment for AAA or lower-extremity aneurysms.

Results also showed significant variation in the strongest recommendation (class I, level of evidence A) between procedures:

  • 24% for endovascular therapy for stroke;
  • 18% for endovascular or surgical revascularization for lower-extremity PAD;
  • 20% for endovascular or surgical treatment for aneurysms of the abdominal aorta and the lower extremities; and
  • 0% for all other peripheral vascular interventions.

The most common recommendation for all peripheral vascular interventions was class II-C (C-‘expert’ opinion) (27%), followed by class II-B(B-Single RCT /multiple observational data) (26%).


Changes over time

From the 2005 to 2011 guidelines, the researchers observed some changes in the total number of recommendations.

For lower-extremity PAD, the number of recommendations decreased from 20 to 11 for endovascular therapy and from 29 to 11 for surgery. There were no increases in recommendations supported by level A evidence for either treatment, but the number of class I indications decreased from 10 to three for endovascular therapy (P = .27) and from 19 to five for surgical revascularization (P = .29).

For endovascular stroke therapy, there were no major changes in the number of recommendations or in level A evidence over time. However, level B evidence increased and level C evidence decreased.

The variation in the guidelines indicates that many recommendations in this area are based on lower quality of evidence or expert opinion.


Editorial Commentary

In an accompanying editorial, David W. Lee, MD, and Matthew A. Cavender, MD, MPH, both from the University of North Carolina at Chapel Hill, echoed the need for better evidence.

Research networks that facilitate comparative effectiveness studies in patients with peripheral vascular disease could help advance the field. Furthermore, the clinical trial infrastructure put in place for ongoing studies such as BEST-CLI and CREST-2 could provide a framework for additional studies in PAD, and multidisciplinary initiatives such as the Pulmonary Embolism Response Team Consortium can help secure funding for high-quality research. The use of existing registries, formulation of pragmatic trials nested in such registries, as well as improving data collection within these registries, could supply important information. The overarching goal of research in this field is to determine which treatments are most effective best on higher quality evidence.



  1. Lee DW, Cavender MA. Guidelines for Peripheral Vascular Disease: Where Is the Evidence? Circulation: Cardiovascular Interventions. 2019;12(1). doi:10.1161/circinterventions.118.007561. Lee DW, et al. Circ Cardiovasc Interv. 2019;doi:10.1161/CIRCINTERVENTIONS.118.007561.
  2. Sardar P, Giri J, Jaff MR, et al. Strength of Evidence Underlying the American Heart Association/American College of Cardiology Guidelines on Endovascular and Surgical Treatment of Peripheral Vascular Disease: Circulation: Cardiovascular Interventions. 2019;12(1). doi:10.1161/circinterventions.118.007244. Sardar P, et al. Circ Cardiovasc Interv. 2019;doi:10.1161/CIRCINTERVENTIONS.118.007244.