Why the study?
Does the presence of right ventricular dysfunction or elevated cardiac biomarkers predict early mortality in patients with acute pulmonary embolism classified as low-risk by clinical scores?
Does the presence of right ventricular dysfunction or elevated cardiac biomarkers predict early mortality in patients with acute pulmonary embolism classified as low-risk by clinical scores?
Answers are grounded in this paper and cite their sources.
This editorial argues that clinical risk scores alone are insufficient for identifying low-risk pulmonary embolism patients suitable for home treatment, advocating for the inclusion of imaging or biomarker assessment of right ventricular function.
This editorial refers to ‘Prognostic value of right ventricular dysfunction or elevated cardiac biomarkers in patients with low-risk pulmonary embolism: a systematic review and meta-analysis’†, by S. Barco et al., on page 902. Acute pulmonary embolism (PE) is potentially fatal due either to a first massive episode or early recurrence; however, if diagnosed early, it can be effectively treated and in most cases cured. PE episodes have a wide spectrum of severity: from asymptomatic, detected incidentally to life-threatening with haemodynamic collapse. Consequently, optimal treatment may range from anticoagulant pills taken at home to thrombolysis, percutaneous catheter interventions, or surgical embolectomy. Each treatment method carries risks of specific complications, and should be selected to match patient characteristics and the severity of the PE episode. The assessment of PE severity has evolved over time, oscillating between clinical evaluation and emerging imaging and laboratory methods. Graduating from medical school 40 years ago with Hurst and Schlant’s The Heart, Arteries and Veins under my arm, I could read that the ‘PE of a size sufficient to produce symptoms has a 20–38% mortality rate’.1 In their ‘bible’ text about human pulmonary circulation, Harris and Heath were more specific in writing that ‘the degree of systemic hypotension is proportional to the gravity of prognosis’.2 Thrombolytic trials introduced a new concept of PE severity. Miller and Walsh indices offered quantitative scores based on the number and calibre of occluded pulmonary arteries and on regional lung perfusion, as assessed at angiography. This anatomical approach resulted in new nomenclature: e.g. a Miller score beyond 20/34 (60%) was termed ‘massive’ PE while that between 41 and 60% was ‘submassive’.3 Similar scores have more recently been derived from computed tomography (CT) angiography. For many years, the terms massive and submassive PE have remained equivocal; some have used them to describe the clot burden, while others have used them to grade clinical or haemodynamic compromise. Finally, in his excellent treatise on Death in Pulmonary Embolism Morpurgo wrote ‘in modern parlance PE is called massive not so much according to its anatomical extent or location as on the basis of hemodynamic criteria’.4 This concept prevailed in the following years. The Management Strategy and Prognosis of Pulmonary Embolism Registry (MAPPET) not only reported a sharp increase in mortality among patients with PE who had systemic hypotension, but also provided its definition. It set the cut-off systolic blood pressure (SBP) at 90 mmHg but also included a persistent (>15 min) drop by > 40 mmHg ‘not caused by new-onset arrhythmia, hypovolaemia, or sepsis’.5 This definition has also been promoted by European Society of Cardiology (ESC) Guidelines on PE.6 The guidelines replaced the term ‘massive’ with ‘high-risk’ to highlight the fact that the severity of PE should be understood as the predicted risk of early mortality related to the degree of haemodynamic compromise and, consequently, to the shrinking right ventricle (RV) haemodynamic reserve. Consequently, emerging non-invasive methods assessing right ventricular function and myocardial injury came under the spotlight. Echocardiographic research resulted in embarras de choix, unfortunately without proposing a unique, universally accepted prognostic sign with a well-validated threshold.7 Paradoxically, CT angiography provided a simple RV/left ventricle (LV) ratio taken from routinely acquired horizontal planes as a useful alternative.8 In the meantime, troponin and natriuretic peptide levels were found to also be of prognostic value in PE. Just like with imaging, despite different cut-offs used, biomarkers proved highly predictive of benign outcome if found to be ‘normal’.9 These findings permitted three-level risk stratification: high, intermediate, and low, as indicated by systemic pressure, and cardiac imaging or biomarkers. However, another concept to define the severity of PE was being developed in parallel. The PE Severity Index (PESI), in addition to heart and respiratory rate, SBP, and oxygen saturation, also accounts for individual characteristics of the patient, which do not result from the index embolic episode but potentially aggravate its course. Weighted scores based on 11 parameters define five levels of risk, from very low to very high (Table 1). However, the authors promoted PESI mainly as a way to identify patients with good outcomes. This term was applied to patients in severity classes I and II (30 day mortality 0–1.6 and 1.7–3.5%, respectively), defined by less than 86 collected PESI points. Of note, neither PESI, nor its simplified (s) sPESI version or the Hestia criteria, which was conceived specifically to identify patients suitable for early discharge or home treatment, included parameters related to RV function. Looking at PESI from some distance, it appears to be a universal score that could be applied to any clinical situation, e.g. pneumonia or stroke, indicating elderly patients with comorbidities and/or disturbed key vital signs as those with worse expected outcome Pulmonary embolism severity index. Out of five validated prognostic classes, classes I and II allow the consideration of a patient with pulmonary embolism as being low-risk for 30 day mortality b.p.m., beats per minute. Pulmonary embolism severity index. Out of five validated prognostic classes, classes I and II allow the consideration of a patient with pulmonary embolism as being low-risk for 30 day mortality b.p.m., beats per minute. Pulmonary embolism management algorithm based on severity assessment. Adapted from author's presentation of key messages from ESC Guidelines at the World Congress of Cardiology, Beijing 2010. CT-angio, CT angiography; echo, echocardiography; interm-risk, intermediate risk; Th-lysis, thrombolysis. With this in mind, the 2008 edition of the ESC Guidelines on acute PE suggested ‘clinical assessment’ scores as the last checkpoint before early discharge or home treatment, which should be considered only for normotensive patients who were free from RV dysfunction at imaging or as suggested by elevated cardiac biomarkers (Take home figure).6 This seemed a good compromise but it did not last long. Validation of (s)PESI and Hestia, and the role that PESI played in the first randomized home vs. hospital treatment trial, has reinforced the position of clinical prognostic scores.10 As a consequence, and not without doubts, the ESC 2014 Guidelines simplified the prognostic assessment of normotensive patients. The management algorithm allowed the classification of patients with PE as low-risk based on their (s)PESI score, i.e. on clinical grounds alone, considering further cardiac testing obsolete. Such a strategy was promoted as safe, saving time and money, and facilitating management decisions in settings where point-of-care cardiac imaging and biomarker testing were not available.11 There were several uncertainties related to this approach. In particular, it was unclear whether there are low-risk patients according to (s)PESI who would still have RV dysfunction, and what the prevalence and prognostic significance of such clinical setting would be. At the time of writing this editorial, we admitted a 43-year-old woman with systemic hypertension who, in addition to age, scored only 20 PESI points for tachycardia of 130 b.p.m., and no points for an SBP of 105 mmHg, with a respiratory rate of 28 and oxygen saturation of 92%. Despite being low-risk according to 63 PESI points, at admission she had N-Terminal pro Brain Natriuretic Peptide (NTproBNP) of 4464 pg/mL, high sensitivity troponin (hs Tn) of 64 ng/mL, and severe RV dysfunction on echocardiography. Should we be concerned about the outcome of such patients? Early recurrence may be more threatening in view of already reduced haemodynamic reserve. An undiagnosed right heart thrombus may embolise to lungs with fatal consequences; one of the early home treatment PE trials crashed on such an iceberg.12 In addition, RV dysfunction could also be suddenly aggravated by a mechanism once known as myocardial hibernation or stunning. Predisposing factors are many: RV is submitted to strenuous exercise combating against suddenly increased afterload. RV coronary perfusion may be reduced with leaking troponin, while catecholamines are in excess due to decreased systemic pressure and anxiety about sudden-onset dyspnoea. If no help appears, the RV may simply refuse to contract. Indeed, electromechanical dissociation is the usual mechanism of death in acute PE.4 McConnell et al. report that a reversible regional akinesia of the midportion of the free RV can be found despite hyperkinesis of the RV apex,13 resembling the ballooning of the LV apex see in Takotsubo cases. While such scenarios are not frequent among patients with reassuring (s)PESI/Hestia scores, they would be particularly painful if occurring after discharge in young patients with otherwise excellent life expectancy. The meta-analysis by Barco et al. in this issue of the European Heart Journal is not reassuring.14 The risk of early mortality in patients classified as low-risk based on clinical scores was increased four to six times in the presence of RV dysfunction or injury as detected by imaging or cardiac biomarkers, respectively. Troponin leak alone was associated with mortality of 3.8%, which is comparable to that of acute myocardial infarction. Similar mortality was reported in some of the prospective trials in patients with PE after early discharge or on home treatment.15 Maybe some of those deaths could be prevented by reversing the sequence of testing in the management algorithm? Whether or not we can help patients at increased risk, suggested by imaging and/or biomarkers, by keeping them in hospital is an open question. Roy et al. reported recently that among patients in PESI classes I/II, hospitalization was associated with higher rates of complications than home treatment.16 Interestingly, this was also true for patients in higher-risk classes III/V. Rather than suggesting that hospitalization actually is harmful, the authors explained that this was likely due the way that decisions regarding hospitalization had been made: they were based on gestalt, which proved more effective in predicting a complicated clinical course of PE than ‘severity criteria, even regrouped in a score such as PESI…’.16 By designing clinical risk scores, we probably still have not reached the Holy Grail of infallibility. Therefore, we should consider borrowing good ideas from the past (2008 ESC Guidelines) before leaping ‘back to the future’ with new recommendations. The exclusion of increased RV/LV ratio just by looking at already performed CT angiography ‘costs’ only 3 min. Point-of-care TnT became a standard in most acutely symptomatic patients. At least one of those tests should be done before moving to clinical scores when considering whether or not to send a patient with PE home, but this is just my personal opinion. We will have to wait some months to see what the future management algorithm based on risk stratification in PE recommended by the consensus of experts working on the ESC Guidelines on PE will be. See you at the ESC Guideline Sessions next August!
No takes yet. Share an insight, caveat, or question.
Adam Torbicki (2019) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: