Rethinking Testing for Pulmonary Embolism Less Is More facilitated by Dr Victoria Brazil

Journal Club Summary

Rethinking Testing for Pulmonary Embolism: Less is More

26 March 2014

Facilitated by Dr Victoria Brazil


Pulmonary embolism (PE) is a serious and potentially fatal medical condition where the main artery of the lung is suddenly blocked, most frequently due to a blood clot that formed in the deep veins of the legs and migrated to the lungs. Today, many PEs are diagnosed and treated early thanks to testing by pulmonary angiographs, ventilation:perfusion nucleotide scanning  or computed tomographic (CT) scanning. However, there are concerns about over-testing and the efficacy of the anti-coagulation therapy administered after diagnosis. Potential harms of PE testing include cancer from imaging, renal damage from contrast and complications of treating patients with positive results. Benefits are not so clear-cut. The bulk of PE testing in outpatient settings is conducted in physiologically normal patients – not just in those presenting with symptoms of PE. The combined harms and benefits of PE testing and treatment in an emergency department (ED) setting have not been evaluated.

Paper presented

Victoria presented an editorial by Newman & Schriger1 who used data from an observational study2 on PE testing in an ED setting to examine how many patients might actually benefit from being tested and diagnosed, and how many would benefit from heparin treatment. In this study, more than 8,000 patients were tested, 500 PEs were detected, and 0.2% of patients died of PE. To calculate an estimated aggregate effect of testing for PE, the authors used a “balance sheet” approach that calculates and juxtaposes patient-oriented benefits and harms.

The results and author’s main conclusions were striking:

  • In this study, testing for PE prevented 6 deaths due to PE and 24 major, non-fatal PE events per 10,000 patients.
  • In comparison, testing for PE also caused 36 deaths and 37 non-fatal major medical harms (renal failure, cancer, major haemorrhage) per 10,000.
  • Based on these calculations, the current model of testing causes roughly 6 times as many deaths as it saves lives.
  • Future emphasis on pursuing and treating PE diagnosis should be on patients with a high risk, i.e. who have physiological symptoms. In the meantime, more research into both therapy and diagnosis, and monitoring of outcomes, should help to reduce unnecessary testing and increase patient benefits.


Our discussion focussed on risk management strategies associated with this problem.

Clearly, GPs and ED physicians don’t want to miss a diagnosis and hold the belief that PEs inevitably cause harm, therefore the tendency to test stems from the wish to avoid missing that diagnosis. With this strategy, physicians cover their own risk, but as this study shows, they may not reduce the risk for the patient – quite the opposite.

Since Virchow’s times, the mindset of clinicians has been “PEs are everywhere and fatal, therefore we must find them”. This, together with new technologies, has led to increased testing and early identification and treatment of PE. But are the PEs we detect today clinically and anatomically the same as those described by Virchow? Probably not.

The calculations presented by Newman & Schriger do not change the fact that PE can be deadly. But as they state in their discussion: “When the goal is to detect all clots, the harms of testing outpace the threat of the disease.”

Autopsy studies and other evidence suggest that PE is overwhelmingly non-fatal and that small clots are both transient and normal.

Does this mean that doctors need to move from individual-based practice to population-based practice? What is an acceptable “miss rate”? Given the legal actions that can follow a missed diagnosis (especially in the U.S.), doctors may find it difficult to get their head around this concept. Perhaps a change in the legal framework is needed too. Doctors can be sued for missed diagnosis, but not for harm caused by overdiagnosis.

To change this practice of overdiagnosis, physicians and the research community should at the very least take a step back, re-evaluate their beliefs, and confront the knowledge gap. We do not know which patients with PE are at risk of fatal consequences and it is not established whether anticoagulants present an effective therapy for PE.

In the meantime, patient education and a decision support tool for this problem could help with shared decision making, although we agreed that patients tend to go with the recommendations of their physician. Hence, educating medical students and new guidelines for current physicians might be the better way forward.


  1. Newman DH & Schriger DL (2011). Rethinking Testing for Pulmonary Embolism: Less is More. Ann Emerg Med. 2011; 57:622-627.
  2. Kline JA, Courtney DM, Kabrhel C, et al. Prospective multicentre evaluation of the pulmonary embolism rule-out criteria. J Thromb Haemost. 2008;6:772-780.

Summary provided by Michele Weber