Category: Pulmonary Embolism

Welcome to Yesterday, Have You Heard of PERC?

I usually like these sorts of articles regarding the yield and utilization of CT pulmonary angiograms.  They’re fun to dissect, useful to marvel at the inefficiency of our usage, and finally to feed my editorial hyperbole.  But, not this time.

This is a retrospective study from the University of Michigan comprising six months of CTPA data from 2013.  These authors reviewed charts on 602 consecutive patients and calculated modified Wells and PERC for each, and describe the appropriateness and yields of various cohorts.

Rather than detail these statistics and outcomes – other than to note their overall yield of 61 positives reported out of 602 scans – I’d rather just focus on the 108 patients scanned who were PERC negative.  PERC has been around since 2004, and it’s been percolating into various guidelines and evidence-based algorithms since.  Hello, it’s 2015: why are almost 20% of CTs at an academic medical center PERC-negative?

The authors state two PERC-negative patients had positive CT findings; given the pretest probability, I wouldn’t be surprised if one or both were ultimately false-positives.  Come on, man.

“CT Pulmonary Angiography: Using Decision Rules in the Emergency Department”
http://www.ncbi.nlm.nih.gov/pubmed/26435116

The Battle for Age-Adjusted D-Dimer

Around these parts, we are fans of the age-adjusted D-dimer.  Jeff Kline proposes their use in his algorithm for the diagnosis of PE.  We embed decision-support in our EHR to encourage their use.  But, this new review from Annals of Emergency Medicine describes its test characteristics in the Kaiser Permanente population – and reports the age-adjusted D-dimer is not infallible.

These authors look retrospectively at 31,094 patients over 50, with a chest- or respiratory-related complaint, for whom a D-dimer was ordered.  14,434 of these patients had a D-dimer above the “customary” level of 500 ng/dL, and clinicians ordered 12,486 imaging studies to evaluate for PE.  Of these, 507 were diagnosed with PE.  This gives a 4.1% yield for CTPA – which, frankly, is disturbingly low – but another topic for another day.

The 500 ng/dL threshold was sensitive for 497 of the 507, while using an age-adjusted D-dimer would have reduced sensitivity to 471 of the 507.  Thus, using an age-adjusted D-dimer in this retrospective cohort may potentially have introduced an additional 26 missed PEs.  The savings, however, amount to 2,924 fewer CTPAs – or, roughly, 100 CTs per missed PE.

The contemporaneous Twitter response:

@EBMgoneWILD @ZackRepEM So age-adjusted D-dimer is dead? 26 misses to save $290K in costs = dead.

— Robert McNamara (@RobertMcNamar12) September 4, 2015

I don’t think so – but questions abound, many of which need be directly addressed by our specialty.  What is an acceptable miss rate for pulmonary embolism?  What is an acceptable miss rate of the pulmonary emboli in this age-adjusted range, just above our prior test threshold?  Does the net harm reduction from reduced testing outweigh the harms of missing those PEs?  Do those PEs convey the same level of morbidity or mortality if the diagnosis is missed or delayed?  How does the radiologic false-positive rate trend for PEs whose D-dimers are just over the test threshold?  And, finally – the age-adjusted D-dimer is not a static construct – would other age-adjustment formulas strike a better balance between sensitivity and specificity?

When all the questions are posed, I believe the summative value shows it reduces physiologic harms from testing, harms from healthcare costs, and harms from false-positives.  But, like everything we do, the age-adjusted D-dimer is still deserving of continued questioning and refinement.

“An Age-Adjusted D-dimer Threshold for Emergency Department Patients With Suspected Pulmonary Embolus: Accuracy and Clinical Implications.”
http://www.ncbi.nlm.nih.gov/pubmed/26320520

Your CTPA is Lies

There are a few moments you pat yourself on the back in Emergency Medicine.  The good save.  Shared decision-making that goes well.  And, the small victory when you’ve utilized an evidence-based pathway for pulmonary embolism, and received positive results for the leviathan of over-utilization and over-diagnosis: the CT pulmonary angiogram.

Well, it’s time to deduct about 1.25 fingers from that pat on the back you give yourself, because, unfortunately, radiology PE overcalls may be more rampant than initially thought.

This is a retrospective, single-center study reviewing a year’s worth of CTPA for pulmonary embolism, a total of 937 studies.  Of the studies included, 174 (18.6%) were initially read as positive.  Then, each positive study was reviewed by a panel of three, specially trained chest radiologists, with their consensus read used as the gold standard for diagnosis.  And so: 45 (25.9%) were subsequently judged to be incorrectly read by the original radiologist – a quarter of positive studies! – with those patients almost certainly consigned to at least short-term anticoagulation as a result.

In a light moment in the discussion, the authors helpfully contribute the following commentary:

Furthermore, many pulmonary CTA examinations in our institution are ordered by the emergency department before assessment by the admitting medical team.

My heart goes out to the poor Scottish EM physicians, for whom their radiology colleagues apparently have quite the low opinion for appropriate testing.  However, the authors’ attention may be better spent further discussing their own false-positive rate, which is double the ~11% rate of other similar reviews.  They also do not provide any accompanying data on the rate of false-negatives, although, in theory, these should be less clinically important.

So, think twice about doing your little happy dance for a positive CT – if your pretest likelihood was low, and the PE is subsegmental, there’s a substantial chance the stars have aligned in just the wrong constellation.

“Overdiagnosis of Pulmonary Embolism by Pulmonary CT Angiography”
http://www.ncbi.nlm.nih.gov/pubmed/26204274

Gestational Age and D-Dimer Levels

In general, the utility of D-dimer for the evaluation of venous thromboembolism declines with gestational age.  The typical cut-offs for the 95th percentile, depending on your assay, become less and less relevant as pregnancy progresses.  Wouldn’t it be nice, perhaps, if we had reliable data?

So, well, here’s something:

One glaring hole in this data is the broad inclusion criteria of “healthy” women.  No testing was specifically performed to exclude asymptomatic venous thromboembolism, so the possibility exists of inclusion of small, subsegmental pulmonary emboli, or of non-occlusive lower extremity deep venous thrombosis.  The effect on this data would be to increase the 95% percentile, and to widen the 95th percentile confidence interval.

Jeff Kline has proposed gradually increasing cut-offs of 750, 1000, and 1250 ng/mL for the first, second, and third trimester, respectively (based on a standard cut-off of 500 ng/mL).  This sample is much larger than the one cited by Kline in his “PE in pregnancy” algorithm, but his appear to be reasonable, sensitive cut-offs.  By far, the most important aspect of evaluating pulmonary embolism in pregnancy is simply to communicate the uncertainty, and to inform and share decision-making with the patient along the way.

“Gestation-specific D-dimer reference ranges: a cross-sectional study”
http://www.ncbi.nlm.nih.gov/pubmed/24828148

Can You Diagnose PE With a Walk Test?

So, no.

You can stop reading now, if that’s enough information to satisfy your curiosity.  There is, however, a little more to it.

These authors describe a prospective evaluation of 114 Emergency Department patients with either suspected or confirmed acute pulmonary embolism.  Patients were enrolled by convenience selection during the hours research assistants were in the ED.  Each enrolled patient underwent a 3-minute walk test while research assistants measured changes in heart rate, respiratory rate, and oxygen saturation.

In short, ambulation induced significant changes in heart rate and oxygen saturation between those who did, and did not, have pulmonary embolism.  A change in heart rate of 10 bpm gave a sensitivity of 97% (95% CI 83 to 99%) and specificity of 31% (95% CI 22-42%), while a drop in O2 saturation of 2% gave a sensitivity of 80% (95% CI 63 to 91%) and specificity of 39% (95% CI 30 to 50%).  Obviously, these test characteristics are poor – excepting, perhaps, a potentially useful negative likelihood ratio, particularly when both variables are utilized.  However, there are also serious issues with their gold-standard for diagnosis of pulmonary embolism – with nearly 30% of their cohort undergoing ventilation/perfusion scans.

I appreciate these authors’ attempt to describe the test characteristics of, essentially, a free, non-invasive physiologic stress – and, even if the current data does not support routine use, it’s probably worth continuing to explore.

“Ambulatory vital signs in the workup of pulmonary embolism using a standardized 3-minute walk test”
http://www.ncbi.nlm.nih.gov/pubmed/26034913

Might High-Sensitivity Troponin Out-Perform PESI?

Risk-stratification of patients diagnosed with acute pulmonary embolism is generally considered a valuable enterprise.  High-risk patients are reasonable to observe as an inpatient for deterioration leading to thrombolysis or other invasive procedures, while low-risk patients can be obviated the costs and risks of an inpatient stay.  The Pulmonary Embolism Severity Index is in use in many settings to make such a determination – calibrated for maximum sensitivity to detect adverse events.

Cardiac troponin has been similarly used for risk-stratification – but mainly for determining “high-risk” and the spectrum of submassive PE, as many patients with negative conventional troponins still progress to poor outcomes.  This study evaluates the utility of the highly-sensitive troponin – threshold of detection 0.012 ng/mL – for risk-stratification.

Based on retrospective review of 298 consecutive patients with acute PE, these authors found about half had a detectable hsTnI, while the remainder were below the limit of detection.  With regards to “hard events” as a primary outcome – death, CPR, or thrombolysis – no patient with an undetectable troponin had such an event in the hospital.  Conversely, 15 (9%) patients with a detectable hsTnI suffered a serious outcome.  Interestingly, based on a rough evaluation of the Kaplan-Meier survival curves, even patients stratified as intermediate, high, or very high risk by PESI were still event-free if their hsTnI were negative – while a small number of patients low-risk by PESI had events, but only with positive hsTnI.

This is not the most robust evaluation of such risk-stratification, and there’s clearly some oddities in the chart review, given an odd spate of low-risk patients deteriorating between days 10 and 20.  However, it may be the case the hsTnI does as-good or better at risk-stratifying than our current tools – and may be considered for inclusion into future ones.

“The Prognostic Value of Undetectable Highly Sensitive Cardiac Troponin I in Patients With Acute Pulmonary Embolism”
http://www.ncbi.nlm.nih.gov/pubmed/25079900

Adverse Outcomes are Very Rare After Pulmonary Embolism

And, using Science! we can clearly see – the sicker the patient, the more likely the poorer outcome.

This is a retrospective evaluation of EINSTEIN PE, using the trial data to retrospectively evaluate the simplified pulmonary embolism severity index (sPESI) score.  The sPESI uses the following features to generate a risk score, one point each:

  • Age > 80 yr
  • History of cancer
  • Chronic cardiopulmonary disease
  • Pulse ≥ 110 beats/min
  • sBP < 100 mm Hg
  • Arterial oxyhemoglobin saturation level < 90%

“Low risk” for sPESI is a score of zero, and supposedly confers a splendid prognosis with regard to subsequent VTE-related complications or all-cause mortality.

EINSTEIN PE randomized 4,832 patients to either oral rivaroxaban or “standard therapy”, typically a parenteral heparinoid followed by warfarin.  This analysis was able to calculate sPESI scores for 4,831 from the trial data, and stratified the outcomes by sPESI 0, 1, and 2.

As you might expect, patients with sPESI of 0 had recurrent VTE, fatal PE, and all-cause mortality rates in the fraction of a percent.  Interestingly, sPESI scores of 1 had similar, tiny fractions of recurrences of VTE and fatal PE, although all-cause mortality and bleeding complications were higher.  sPESI of ≥2, showed significant divergent outcomes both early and throughout the treatment period – although, again, the VTE-related and treatment-related morbidity and mortality remained near 1% in the first 30 days. All-cause mortality was much higher, however, over 10% during the treatment period – which makes sense, considering these patients have significant physiologic derangements, along with the underlying disease process responsible for inciting a symptomatic pulmonary embolism.

But, even with sPESI ≥2, the absolute risk for VTE-related morbidity and mortality was less than 2%.  It is probably reasonable to continue questioning what outcomes advantage can be attributed to initial hospitalization, and whether otherwise appropriate patients might yet be candidates for outpatient therapy.  With such low absolute rates of poor outcomes, it may be difficult to detect a difference attributable to a management effect.

“Treatment of Pulmonary Embolism With Rivaroxaban: Outcomes by Simplified Pulmonary Embolism Severity Index Score from a Post Hoc Analysis of the EINSTEIN PE Study”
http://www.ncbi.nlm.nih.gov/pubmed/25716463

Using Patient-Similarity to Predict Pulmonary Embolism

Topological data analysis is one of the many “big data” buzzphrases being thrown about, with roots in non-parametric statistical analysis, and promoted by the Palo Alto startup, Ayasdi.  I’ve done a little experimentation with it, and used it mostly to show the underlying clustering and heterogeneity of the PECARN TBI data set.  My ultimate hypothesis, based on these findings, would be that patient-similarity is a more useful predictor of individual patient risk than the partition analysis used in the original PECARN model.  This technique is similar to the “attribute matching” demonstrated by Jeff Kline in Annals, but of much greater granularity and sophistication.

So, I should be excited to see this paper – using the TDA output to train a neural network classifier for suspected pulmonary embolism.  Using 152 patients, 101 of which were diagnosed with PE, the authors develop a topological network with clustered distributions of diseased and non-diseased individuals, and compare the output from this network to the Wells and Revised Geneva Scores.

The AUC for the neural network was 0.8911, for Wells was 0.74, and Revised Geneva was 0.55. And this sounds fabulous – until it’s noted the neural network is being derived and tested on the same, tiny sample.  There’s no validation set, and, given such a small sample, the likelihood of overfitting is substantial.  I expect performance will degrade substantially when applied to other data sets.

However, even simply as scientific curiosity – I hope to see further testing and refinement of potentially greater value.

“Using Topological Data Analysis for diagnosis pulmonary embolism”
http://arxiv.org/abs/1409.5020
http://www.ayasdi.com/_downloads/A_Data_Driven_Clinical_Predictive_Rule_CPR_for_Pulmonary_Embolism.pdf

Dueling PE Meta-Analyses

A guest post by Rory Spiegel (@EMNerd_) who blogs on nihilism and the art of doing nothing at emnerd.com.

Nothing sparks controversy quite like a discussion on the utility of thrombolytics. No sooner had the wave of debate brought on by the publication of the PEITHO trial and its finding of no overall mortality benefit died down, did JAMA stoke these flames with the publication of a meta-analysis including the entirety of the literature on thrombolytic use for pulmonary embolism. Examining 16 trials, the authors found a statistically significant absolute mortality benefit of 1.12% or an NNT of 59 patients. This benefit was offset by the increase in major bleeding events observed in those given thrombolytics (9.24% vs 3.42%) with a 1.27% absolute increase in ICH.



The cascade of incendiary events continued when one week later a second meta-analysis examining the very same question was published in Journal of Thrombosis and Haemostasis. Only these authors claimed to find the exact opposite of their JAMA counterparts. In this case the authors found no statistical improvement in mortality in the patients given thrombolytics when compared to those given a placebo. Despite these contradictory claims, a more comprehensive inspection reveals these meta-analyses are extensively saying the same thing. A comparison of the two serves as a timely reminder that conclusions reached from any meta-analysis is primarily dependent on the trials selected for inclusion.  The JAMA meta-analysis included 2115 patients in 16 trials, while the Journal of Thrombosis and Haemostasis examined only 1510 patients in 6 trials. Interestingly, the absolute risk reduction in mortality was 1.12% in the JAMA publication vs 1.4% in the Journal of Thrombosis and Haemostasis publication. Though the JAMA analysis had an overall smaller absolute risk reduction, the result reached statistical significance due to the larger sample size.



More importantly the results of these publications should not come as a surprise. Before the publication of PEITHO the data on thrombolytics for PE was sparse. Most of the trials suffered from small sample sizes and questionable methodology. It is the amalgamation of these small trials that accounts for the mortality benefit in both meta-analyses. In the JAMA publication the mortality difference consisted of 17 fewer deaths in the thrombolytic arm compared to the placebo. All of which originated from these small underpowered studies. Conversely, the two large high quality trials (PEITHO and MSPPE) consisting of 1005 and 256 patients respectively made up the majority of patients meta-analyzed, neither of which found a mortality benefit with the use of thrombolytics. Moreover the 2% absolute increase in ICH seen in the PEITHO cohort is only diluted by the inclusion of these small trials, whose sample sizes were not powered to detect such rare events. A more elegant design would be to utilize a weighted average or one of the various statistical methods that takes into account each study’s sample size and event rate, allocating a greater weight to the hardier cohorts. Though one might argue an equally elegant solution would be to not include such flawed trials in the first place.


The publication of these dueling meta-analyses highlights the flaws of such statistical endeavors. Small trials with flawed methodological designs are prone to fall victim to publication bias and the fluctuating whims of chance. Collecting this data and attaching a statistical judgment to it does not correct these imperfections, it augments them. The overall benefit of thrombolytics in PE is yet undetermined, but the answer will not be elucidated in such analysis. We require further large randomized controlled trials like the PEITHO trial. Adding small flawed cohorts to this robust dataset does nothing but muddy the already murky waters.

“Thrombolysis for pulmonary embolism and risk of all-cause mortality, major bleeding, and intracranial hemorrhage: a meta-analysis.”
http://www.ncbi.nlm.nih.gov/pubmed/24938564

“Impact of the efficacy of thrombolytic therapy on the mortality of patients with acute submassive pulmonary embolism: a meta-analysis.”
http://www.ncbi.nlm.nih.gov/pubmed/24829097

Damnit, Who Ordered That D-dimer?!

We live in strange, complicated times.  Popular to our twisted reality are haphazard panels of cardiac biomarkers, ordered unthinkingly via triage protocol or unwittingly by physicians using order sets.  Troponins, myoglobin, creatinine kinase, brain naturetic peptide, and, sometimes, D-dimer results will arrive on patients for whom no suspicion of cardiovascular disease is present.

So, what do you do with that positive D-dimer in a patient who, until just that moment, appeared to be zero-risk for pulmonary embolism – possibly, say, by PERC?

This retrospective chart review from four French hospitals identified all patients undergoing D-dimer testing as part of evaluation for pulmonary embolism.  Of 2,791 patients screened with complete data, 1,070 were PERC negative.  Of these 1,070 minimal risk patients, 167 had positive D-dimer.  153 of these 167 underwent diagnostic imaging for PE, with 5 detected.  Therefore, in this cohort, a patient who was PERC negative with a positive D-dimer had approximately 3.0% incidence of PE.

This result is, however, absent any other abstracted objective risk-stratification.  PERC was designed to work in concert with other objective or gestalt risk-stratification into a low-risk cohort.  So, even though these authors claim a number of unnecessary imaging studies, it is likely a handful of these were reasonable tests utilizing risk factors outside of PERC.

Regardless, please carry on properly ignoring the majority of inadvertent positive D-dimers – if PE is not reasonably in the differential, as it was in this study, the prevalence of PE will still be vanishingly small.

“Pulmonary Embolism Rule-out Criteria vs D-dimer testing in low-risk patients for the diagnosis of pulmonary embolism: a retrospective study in Paris, France.”
http://www.ncbi.nlm.nih.gov/pubmed/24736129