(Last Updated On: May 3, 2018)

Simplified Diagnostic Management of Suspected Pulmonary Embolism (the YEARS study): a Prospective, Multicentre, Cohort Study.

van der Hulle, Tom, et al. The Lancet 2017 . [paper]

Why I chose this study

Determining if a patient has a pulmonary embolism (PE) is a routine occurrence in the emergency department (ED) and on the medicine inpatient service. I chose this article after a discussion with an attending regarding ruling out PE with clinical criteria, while one of our patients was pending a D-dimer for observation vs full inpatient admission.

Background

Currently our PE work up algorithms include Well’s criteria, Geneva scores, PERC and rolling the dice on a D-dimer before ordering pulmonary CT angiography (CT-PE). However, we still order lots of CT-PEs that we suspect will be negative after having that elevated D-dimer come back. Knowing that we accept a ~1.8% miss rate (pre-test probability of PERC rule out and miss rate of CT-PE), is there any way we can reduce our CT-PE use further?

Research question

Can the YEARS clinical decision tool be used for ruling out a PE?

YEARS clinical decision tool, fig 1 from van der Hulle et al 2017

 

Study Design

This was a prospective cohort study of patients with suspected PE, identified by attending physicians, at 12 hospitals across the Netherlands from Oct 5, 2013 to July 9, 2015. It needed 3260 patients for power of alpha of 5% and assuming 7.5% loss of follow up. The YEARS clinical tool was applied to patients and CT-PE studies were ordered according to the decision tool.

Follow up was done either having patients return for symptoms of VTE or calls done at 3 months to assess for symptoms of VTE.

Inclusion

  • Patients ≥ 18 years old with suspected first or recurrent PE.

Exclusion

  • Treatment with anticoagulation (AC) started 24 hrs or more prior.
  • Life expectancy < 3 months.
  • Geographic inaccessibility precluding follow up.
  • Pregnancy.
  • Contraindication to CT-PE – Allergy to IV contrast dye, renal insufficiency.

Primary Outcome

3 month incidence of symptomatic venous thromboembolism (VTE). Diagnosis made via CT, pulmonary angiography, high probability on VQ scan, ultrasound, or death where PE was a possible cause.

Secondary Outcome

Number of CTs compared to two level Well’s criteria with a dimer < 500 or an age adjusted dimer. Post hoc analysis for comparison.

Analysis

Analysis done with a per-protocol approach for the primary outcome and an intention-to-diagnose approach for the secondary outcome.

Results

3465 patients enrolled after exclusions. Overall PE rate was 13% (456) which is higher than the US and lower than Europe. 18 patients (0.61%) considered ruled out either with decision tool or with CT-PE were diagnosed or suspected to have VTE at 3 months. If you assume worst case scenario and the patients lost to follow up all had thromboembolism the number increases to 23 patients total (0.78%).

  • 1651 ruled out with YEARS decisional tool
    • 40 patients crossed over and received CT-PE showing 3 PEs. Considered failures in intention-to-diagnose protocol.
    • 18 treated with anticoagulation anyway due to other reasons (atrial fibrillation, superficial thrombophlebitis, pulmonary hypertension, etc).
    • 7 patients returned with symptoms of VTE (includes the 3 above).
    • 4 lost to follow up.
  • 1358 ruled out after CT-PE
    • 40 treated with anticoagulation anyway.
    • 11 with VTE on follow up at 3 months.
    • 1 lost to follow up.
  • 1174 (34%) would have been ruled out with Well’s and 500 ng/mL D-dimer (applied post hoc).
  • 1348 (39%) would have been ruled out with Well’s and age adjusted D-dimer.

Out of the patients excluded from CT-PE by applying the YEARS clinical tool seven patients (0.43%) had a future diagnosis of VTE. Three were found on CT when protocol wasn’t followed, of note all three had moderate to high risk via Well’s with low D-dimer. At least two of these seven patients may have not been true misses based on other comorbid conditions or lack of definitive diagnosis.

Use of the YEARS criteria reduced use of CT-PE scans by 14% vs Well’s with D-dimer and 9% vs Well’s and age adjusted D-dimer.

Resin cast of pulmonary arteries

Discussion

The YEARS criteria is a new tool that may still need further validation, but does show another way to significantly reduce the total number of CT’s in a large population with a D-dimer limit (1000 or 500) that is based on pre-test probability. Initially I was skeptical of the study as it didn’t account for malignancy, long travel, etc in their decision tool like the Well’s score does, but looking at the data, these patients were all distributed in the population. On subgroup analysis the decision tool did well in these groups too. This clinical decision tool is simple to do in the ER as well. Three values obtained from history and exam and one lab value instead of complicated scores with point systems.

Strengths of the study – large population, data matches up with prior PE studies, analysis across multiple subgroups also validates findings.

Weakness – not a true randomized control, but post hoc analysis is very much appropriate. Younger population.

Only 0.43% missed VTE is actually really good considering what we accept with PERC. Looking at the 7 missed cases, I personally would have scanned patient 2,5,6,7 and possibly 4, either due to a high Well’s or elevated age adjusted D-dimer. This is the limitation with decision tools or D-dimer values, we can’t use them blindly, if clinical gestalt is high enough then the more specific test should be ordered. I do think that this provides more evidence that a one-value (positive or negative based on a cut-off of 500 ng/ml) D-dimer might not be the best thing and shows evidence that we still have a ways to go to reduce CT burden.

I would really have like to see the characteristics of the 40 patients where the protocol wasn’t followed and if all of these cases had much higher Well’s scores (similar to the 3 who had PEs). It would also be interesting to see how many VTEs would have been missed if only the Well’s and D-dimer or age adjusted D-dimer was used. In addition, I wonder how many had YEARS score ≥2 but ruled out with a dimer <500 ng/mL.

Is this score ready for prime time and will it change my current practice? Probably not yet, but I’m excited to see further validation and possibly a randomized controlled trial comparing YEARS to current practice.

 

Vinay Mikkilineni is a EM/IM PGY5, he will be staring a global health fellowship at UIC in the fall.

 

 

 


EM/IM Sessions are reviewed in journal club style by the current attendings and residents, as well as alumni of the UIC IM/EM program prior to publication. This post was specifically reviewed by Adam Rodos, MD, IM/EM Assistant Program Director. Elspeth Pearce, MD Editor.