Incidence of pulmonary embolism in emergency department

Prevalence of Pulmonary Embolism among Patients Hospitalized for Syncope¹

Clinical Question

• In adult patients who are hospitalised for a first episode of syncope, what is the prevalence of pulmonary embolism (PE)?
• The authors were concerned that PE is an underappreciated cause of syncope

Design PESIT (Pulmonary Embolism in Syncope Italian Trial)

• Multi centre cross-sectional study
• Estimated sample size of 550 patients to obtain a two sided 95% confidence interval of 2.5% (on the basis of the prevalence in a pilot study)

Setting

• From March 2012 until October 2014
• Eleven general hospitals in Italy

Population

• Inclusion:
  • Adult patients > 18 years old
  • Presenting to the emergency department with a first episode of syncope AND admitted to a medical ward
  • “Syncope” defined as a full loss of consciousness for less than 1 minute followed by spontaneous complete resolution 
• Exclusion:
  • On anticoagulation treatment
  • Recurrent syncope
  • Obvious causes for syncope such as epileptic seizures, stroke and head trauma

• 2584 patients were assessed for eligibility; 560 patients were included

Intervention/Work up

• Standardised diagnostic work-up of syncope based on international guidelines²
  • Evaluation to define a subgroup with a potential alternative explanation for syncope
• Diagnostic work-up for possible PE based on the algorithm of the 2014 guidelines of the European Society of Cardiology ³, using:
  • Pre-test clinical probability - simplified Well’s score (PE unlikely vs likely)
  • D dimer assay if PE unlikely according to Well’s criteria
• If “unlikely” pre-test probability AND negative D-dimer assay: no further work up
• If “likely” pre-test probability AND/OR positive D-dimer assay: further work up
  • Computed tomography pulmonary angiography (CTPA)
  • OR ventilation perfusion lung scan (V/Q scan) (if renal impairment or contrast allergy)
• Evaluation of thrombotic burden in patients with PE through
  • Identification of the most proximal location of embolus on CTPA
  • Measurement of the severity of the perfusion defect on V/Q scan
    
    

Statistics

• Prevalence of pulmonary embolism and associated 95% CI in (sub)group
• Comparison of baseline characteristics in PE and non-PE group
  • Chi square test for categorical variables
  • Student’s t test for continuous variables
• Linear regression for Odds ratios with 95% CI

Outcome/Observations

• Prevalence of PE in patients hospitalised for first episode of syncope
  • Entire cohort: 560 patients were included in the study, 230 underwent imaging or autopsy, and 97 received a diagnosis of PE or thus 17.3% (95%CI 14.2 – 20.5)
  • Cohort with high pre-test probability AND/OR positive D-dimer: 42.2% (95%CI 35.8 – 48.6)
  • Cohort with “syncope of undetermined origin”: 25.4% (95% CI 19.4 – 31.3)
  • Cohort with “potential alternative explanation for syncope”: 12.7% (95% CI 9.2-16.1)
• PE ruled out in 58.9% of included patients based on low pre-test probability and negative D dimer assay
• Among patients with PE diagnosed with CTPA, the most proximal location of embolus was:
  • Main pulmonary artery: 41,7%
  • Lobar artery: 25%
  • Segmental artery: 26.4%
  • Subsegmental artery: 6.9%
• Among patients with PE diagnosed with V/Q scan, the perfusion defect (as compared to the total lung area) involved
  • More than 50%: 16.7%
  • Between 26 and 50%: 33.3%
  • Between 1 and 25%: 50%

Authors’ Conclusions

• PE was confirmed in 1/6 (17%) of adult patients hospitalised for a first episode of syncope, not on anticoagulant treatment
• Although an even higher PE prevalence of 25% was noted in the subgroup with syncope of undetermined origin, also almost 13% of patients considered having another clinical explanation for the syncope had PE.
• Syncope due to PE is thought to occur because of a sudden obstruction of the most proximal pulmonary arteries that lead to a transient depression in cardiac output. However, in approximately 40 % of the confirmed PE cases the vascular obstruction was “small” (<25% perfusion defect on VQ scan or most proximal location of embolus in (sub)segmental artery). This suggest there might be other associated mechanisms responsible for the syncope.

Strengths

• Work up for PE regardless of whether another explanation was suggested clinically
• Multiple centres with consistent results
• Subgroup analyses were planned a priori

Weaknesses

• First, Study period quite long (2.5 years) and as such clinical practice may have changed impacting the results, and hence having introduced a bias.
• Second, No inclusion of ambulatory/discharged patients
• Third, Syncope diagnosis based on history of patient/bystanders
• Fourth, Study results not applicable in patients on anticoagulation treatment or with multiple syncope
• Fifth, The PE incidence number of 17% is eye-catching because it’s entirely discordant with the 2 to 3 percent prevalence of PE in syncope from previously published literature^4,5 The authors claim that their number is more accurate but have no solid base for these assumptions because:
  • The fact that 1867 out of the 2584 patients with diagnosis of syncope were discharged to home introduces a large bias as we don’t know how many of those would fit a PE diagnosis
  • Therefore, at most we can conclude that the PE incidence is 97 out of 2584 or thus 3.8% which is closer to the previously reported ranges.
• Sixth, there are some very interesting differences in the baseline demographics of the PE confirmed and no PE group.
  • 73 patients out of 97 had clinical signs and symptoms present (signs of deep venous thrombosis, previous DVT or PE, tachycardia, tachypnea,…) suggesting that PE diagnosis was simply missed by the emergency physicians
• Seventh, the results need to be interpreted with caution as previous data has shown that the use of CTPA may overestimate PE in distal branches with about 25%⁶.
• Eight, this study appears to be an example of over diagnosis but is being interpreted as a call to change practice and as such should be interpreted with CAUTION, as the (bleeding) risks of (unnecessary) administration of oral anticoagulation in a fragile elderly population may outweigh possible benefits
• Ninth, the @NEJM twitter handle spread the news that 1 in every 6 patients with syncope or thus 17% has PE (Figure) and thus stimulating possible overuse of anticoagulants in a fragile elderly population without critical analysis and tapering of the conclusions (see caution above)

Take Home Messages

• In patients with a first episode of syncope, even in the subgroup with a potential other explanation, PE should always be in your differential diagnosis and excluded if considered a possible cause and there is a high pre-test probability and/or a positive D-dimer assay in combination with vital signs abnormalities and clinical signs suspicious of DVT. This is nothing new as nicely stated by Rory Spiegel⁷ “Furthermore, it is important to note, this is not a cohort of 97 pulmonary embolisms in 560 patients as it will inevitably be portrayed. Rather this was 97 (3.8%) radiographic pulmonary embolisms in 2584 patients presenting to the Emergency Medicine for a syncopal event.”
• The results of this study however cannot be generalised to other emergency departments where a good clinical and physiological assessment of the patient should normally identify PE much earlier in the diagnostic process.
• The study does not support the statement that “smaller” PE with limited obstruction can be a potential cause of syncope^8,9
• There is no indication that this study has identified a patient population with occult PE.

References and useful links

1. Paolo Prandoni, M.D., Ph.D., Anthonie W.A. Lensing, M.D., Ph.D., Martin H. Prins, M.D., Ph.D., Maurizio Ciammaichella, M.D., Marica Perlati, M.D., Nicola Mumoli, M.D., Eugenio Bucherini, M.D., Adriana Visonà, M.D., Carlo Bova, M.D., Davide Imberti, M.D., Stefano Campostrini, Ph.D., and Sofia Barbar, M.D., for the PESIT Investigators^* Prevalence of Pulmonary Embolism among Patients Hospitalized for Syncope. N Engl J Med 2016; 375:1524-1531 View on PubMed - Download PDF
   Trial registration number: NCT01797289
   PMID: 27797317 DOI: 10.1056/NEJMoa1602172
2. Guidelines for the diagnosis and management of syncope (version 2009). Eur Heart J 2009; 30: 2631-71. http://eurheartj.oxfordjournals.org/content/ehj/30/21/2631.full.pdf
3. ESC Guidelines on the diagnosis and management of acute pulmonary embolism, 2014 http://eurheartj.oxfordjournals.org/content/early/2014/08/28/eurheartj.ehu283
4. Cook OG, Mukarram MA, Rahman OM, et al. Reasons for hospitalization among emergency department patients with syncope. Acad Emerg Med. July 18, 2016; [ePub ahead of print]
5. Blog article: Pulmonary embolism prevalence examined in patients with syncope http://www.acepnow.com/article/pulmonary-embolism-prevalence-examined-patients-syncope/ posted by Ryan Patrick Radecki, MD, MS on December 14, 2016.
6. Hutchinson BD, Navin P, Marom EM, et al. Overdiagnosis of pulmonary embolism by pulmonary CT angiography. AJR Am J Roentgenol. 2015;205:271-277.
7. Blog article: EM Nerd-The Case of the Incidental Bystander. http://emcrit.org/emnerd/the-case-of-the-incidental-bystander/ posted by Rory Spiegel on October 20, 2016.
8. Blog article: PESIT Investigators: the incidence of PE in those hospitalized following first syncope. pulmccm.org. http://pulmccm.org/main/2016/n-engl-j-med-review/pesit-investigators-pulmonary-embolus-incidence-hospitalized-following-first-syncope/. Published October 2016.
9. Blog article: JC: Prevalence of PE in patients with syncope. St.Emlyn’s. http://stemlynsblog.org/prevalence-of-pe-in-patients-with-syncope-st-emlyns/ posted by Simon Carley on October 20, 2016.