In this article the Authors from Chicago discuss how the concept of PE has changed over the last 100 years. Originally PE was diagnosed clinically with minimal or no testing, so the patients had to be obviously unwell, short of breath, coughing up blood with abnormal ECG and so on. These patients as one might expect had mortalities of over 30%. The introduction of anticoagulation in the 1930’s was done after one small trial, which would not cut the mustard, by today’s standards. The mortality of PE continued to fall to 3% over the intervening years. A large part in that fall was the fact more PE was being diagnosed radiologically or at a lower threshold.
The introduction of the CTPA and more and more sensitive scanners has led to the diagnosis of smaller and smaller PE’s, which likely have no clinical implications for the patient. This is over diagnosis. Using national all cause mortality data bases the incidence of PE before 1990 was 62/100k and post CT has doubled to 112/100k whilst the mortality has been almost exactly the same (12.3 to 11.9/100k).
The last issue discussed is does anticoagulation help? They argue probably for large PE and DVT but the size of the effect or what end points are helped are not clear. Then they point out that studies of silent PE diagnosed after DVT found anticoagulation did not alter mortality or hasten resolution of VQ changes.
Stein P., Goodman L., et al Diagnosis and management of Sub segmental Pulmonary Embolism: Review and Assessment of the Options Clinical and Applied Thrombosis/Haemostasis 18(1) 20-26 2012
These authors review the concept of isolated sub segmental PE and suggest that it NOT be treated if the patient has
- Good Pulmonary reserve
- No DVT on serial tests
- A transient reversible risk factor
- No history of CVC or AF
- Compliant and trustworthy patient.
Before accepting this recommendation bear in mind no RCT of this strategy exist therefore it should be done with due caution.
Isolated sub segmental PE are found only in a sub segmental branch but no larger vessels and using primitive 4 slice CT were found in 4% of PE patients. Incidental PE is found in patients undergoing CT for non-PE reasons leading to the suggestion that one of the normal functions of the lung is to clear thrombi and prevent them entering the arterial circulation. From PIOPED II more powerful CT scanners were more likely to detect sub segmental PE but DVT was usually absent if the PE was limited to the Sub segmental regions. The authors don’t mention the fact that CTPA is about 96% specific, meaning a false positive rate of 4%, hello ??? are these the isolated unexpected sub segmental PE they are talking about???
This leads to the risks of anticoagulation. Meta analysis of 10K patients treated with VitK antagonists for PE, suggests a major bleeding rate of 7% and a death rate of 1.3% annually.
These risks of anticoagulation then need to be weighed against the benefits. The evidence of safety in not treating only comes from retrospective analysis and this can be misleading, as those who weren’t treated were so outrageously low risk that no one thought they even had a PE, thus making the treatment safer than it is. As most PE come from large vein DVT it is argued that if there is no DVT on serial examination of the legs and VQ scans are non diagnostic there is a very low risk of PE in follow up.
There is an ongoing prospective observational study of withholding anticoagulation in sub segmental PE with no DVT http://clinicaltrials.gov/ct2/show/NCT01455818.
So what can we take from this? Some changes that look like PE on CTPA may not be clinically important PE’s and there are risks of anticoagulation that may outweigh any benefits of treatment. Before embarking on even investigating patients for PE we should try and make a proper clinical assessment of pre-test probability of disease otherwise we may find non disease (false positives) or over diagnose (true positives but not clinically meaningful) the patients exposing them to all the risks of treatment with none of the benefits.