Category: Radiology

Clearing the Cervical Spine with Distracting Injuries

Ah, the “distracting” injury. An utterly subjective and modifiable component of cervical spine clearance in the NEXUS criteria. Is it an isolated finger dislocation? Is it a femur fracture? We’ve all seen patients writhing or stoic in the face of either. And, then, factor in any prehospital analgesia ….

This is a prospective, observational study coming out of the American Association for the Surgery of Trauma evaluating the effects of distracting injury on cervical spine clearance. For their purposes, the following injuries were considered “distracting”:

Skull fracture, >2 facial bone fractures, mandible fracture, intracranial hemorrhage (including subdural hematoma, epidural hematoma, subarachnoid hemorrhage, intraventricular hemorrhage, intraparenchymal hematoma), >2 rib fractures, clavicle fracture, sternal fracture, pelvic fracture, thoracolumbar spine fracture, intra-abdominal injury (including solid organ injury, hollow viscus injury, or diaphragmatic injury), femur fracture, tibia/fibula fracture, humerus fracture, radius/ulna fracture, and hip or shoulder dislocation.

The physical exam consisted of midline neck palpation and, absent any contraindication, active range of motion of the neck in flexion, extension and rotation. The cervical collar could be removed at the discretion of the treating team, but – in classic traumatology fashion – all patients underwent CT of the cervical spine, regardless of exam.

There were 2,929 blunt trauma patients with GCS ≥14, and 222 had cervical spine injuries identified on CT. Of these injuries, 25 were “missed” by the clinical exam. The “good news”: the rate of miss was “the same”, regardless of distracting injury – 0.7% vs. 1.3%. The bad news, of course, is that a normal physical examination missed 11% of cervical spine injuries. One patient whose injury would have otherwise been missed by a negative physical examination underwent operative intervention.

While there is some obvious spectrum bias associated with any observational cohort enrolled at trauma centers, it is still a reasonable estimate of the sensitivity and specificity of the physical examination. Clearly, it’s not bulletproof in the context of multi-system trauma – but, depending on the pretest likelihood of a cervical spine injury based on other presenting features, a distracting injury need not disqualify a patient from clinical clearance.

“Clearing the Cervical Spine in Patients with Distracting Injuries: An AAST Multi-Institutional Trial”

https://journals.lww.com/jtrauma/Abstract/publishahead/Clearing_the_Cervical_Spine_in_Patients_with.98558.aspx

Minor Head Injury and Anticoagulants

Guidelines advise performing imaging in those patients on anticoagulants who have suffered minor head injury. We virtually all dutifully obey, because, even though the incidence of intracranial hemorrhage is low – it’s still much higher than zero. But, how high, really? Particularly when they’re sitting there, looking normal, with a GCS of 15?

This systematic review and meta-analysis gathered together 5 studies comprising 4,080 anticoagulated patients with GCS 15 following a head injury. Three of the studies mandated imaging, while the others allowed physician discretion with observation, telephone, and chart-review follow-up to ascertain outcomes. The vast majority of patients were on Vitamin K antagonists, and most mechanisms of injury – where documented – were falls.

Overall, there were 209 (5%) patients with ICH after their fall, nearly all of which were diagnosed at the index visit. There was a wide range of findings, ranging from 4% in the largest studies to 22% in the smaller. However, the larger studies were the ones with the least-complete follow-up after the index event. Therefore, these authors’ random effects analysis and sensitivity analysis generated higher estimates of the incidence, up to 10.9%.

So, while yield is low, we’re still far from having a strategy to support selective scanning to improve value. While it is unlikely many of these would have neurosurgical intervention indicated, a substantial portion likely underwent anticoagulation reversal to prevent further morbidity or mortality. While resource stewardship is always an important consideration, it is unlikely we will anytime soon be altering our approach to minor head injury in the context of anticoagulation.

“Incidence of intracranial bleeding in anticoagulated patients with minor head injury: a systematic review and meta-analysis of prospective studies”

https://www.ncbi.nlm.nih.gov/pubmed/30028001

The Futility of Repeat Imaging in Seizure

In the adult patient with new-onset seizure, it can be reasonable to pursue emergency neuroimaging in many cases. However, the vast majority of presentations to the Emergency Department for seizure are for those with known seizure disorders. In this population, the calculus is different.

This is a retrospective review of 822 presentations for non-index seizures from two hospitals, examining the rate of neuroimaging and incidence of clinically important new findings. Of these, neuroimaging was obtained in about half. Of these 381, only 8 had true positive, clinically important findings on imaging. All of these cases had persistent altered level of consciousness, head trauma, or a focal finding on examination. Absent these factors, there were no cases of true positive imaging findings related to the acute presentation. If imaging were deferred in those cases absent any of those three factors, approximately half of scans could have been obviated.

This is a small sample at two academic institutions and a retrospective evaluation, so it is hardly definitive. However, the authors’ conclusion is reasonable there is certainly an opportunity to further reduce unnecessary imaging in non-index seizures.

“Emergency department neuroimaging for epileptic seizures”

https://www.ncbi.nlm.nih.gov/pubmed/30019464

Again With the Failings of CTPA

Most of the unhinged ramblings on this blog involve lamenting the excessive sensitivity of CT pulmonary angiograms for the diagnosis of pulmonary embolism. “Excessively sensitive for PEs of uncertain clinical signifiance!” and “Too many false positives in an inappropriately selected population!” gloomily drones the author (We can’t all, and some of us don’t).

Now, again, come the baffling attacks from the right – the CTPA isn’t sensitive enough:

“The negative predictive value of CTPA for VTE varies according to pretest prevalence of PE, and is likely to be insufficient to safely rule out VTE as a stand-alone diagnostic test amongst patients at the highest pretest probability of VTE. Prospective studies are required to validate the appropriate diagnostic algorithm for this subgroup of patients.”

Foundational quibbles in the narrative induced by their meta-analysis:

  • Sure, maybe, in the cohort of studies before 2006 – but since then, the number of VTE “missed” by CTPA is less than 1%.
  • A VTE “missed” by CTPA includes lower extremity DVT concurrently diagnosed by duplex ultrasound.  Whether a CTPA should be impugned for failing to include the legs is a separate debate regarding the adequacy of its Natural State of Being.
  • Again, a VTE “missed” by CTPA includes all VTE (including LE DVT) diagnosed in the three-month follow-up period, a timeframe certainly adequate for individuals at high-risk for VTE to develop thrombosis anew.
  • “High” clinical probability in this study refers to those patients with ≥40% pretest likelihood of disease, which is tremendously infrequently encountered in clinical practice.

Clearly, these authors are far from convincing me CTPA is guilty of relevant concerns for inadequate sensitivity in these modern times.  One problem at a time, please; queue up, now.

“Outcomes following a negative computed tomography pulmonary angiography according to pulmonary embolism prevalence: a meta-analysis of the management outcome studies”
https://www.ncbi.nlm.nih.gov/pubmed/29645405

Comparing Unnecessary ED Chest Pain Testing

Since my last post regarding the disutility of coronary CT angiograms for the evaluation of acute chest pain in the Emergency Department was so popular, here’s more: a randomized trial testing CCTA versus stress echocardiogram.

The problem: no specific functional or anatomic testing is routinely necessary in the ED.

And, hidden in this comparison are broad results typical of testing in a “low to intermediate” population with a Diamond-Forrester pre-test probability of 28% and a TIMI score of 0 to 1. Effectively, it’s unimportant to describe their comparison because the incidence of their safety outcomes over a median follow-up of 733 days is so low its virtually impossible to detect a difference. Any Major Adverse Cardiovascular Event occurred in only 4.5% of all patients – but, restricted to the endpoints relevant to the testing performed, only 3.0% had a nonfatal myocardial infarction or cardiac arrest. Any differences in throughput and resource utilization between arms will be related to specific process and protocol implementation unique to the trial institution, so even their findings with respect to their primary outcome are not likely to be generalizable.

But, back to the futility of any test – only 10% of those enrolled were referred for cardiac catheterization, and only half of those received an intervention as a result. Another 10% or so received new or increased pharmacotherapy – likely overlapping with the population undergoing catheterization. Therefore, probably 85% of patients enrolled clearly received no specific benefit from these tests, 5% probably had some benefit, 5% were harmed (excess revascularization, serious complications), and 5% are equivocal.

And, all of this on the foundational premise these tests need to be performed widely, and in the ED. True disease is rare in this population, and the timeliness of diagnosis of CAD does not need to be made at the index visit. This is not high-value medical care.

“Coronary Computed Tomography Angiography Versus Stress Echocardiography in Acute Chest Pain: A Randomized Controlled Trial”

https://www.ncbi.nlm.nih.gov/pubmed/29909113

And Here’s Why CTCA Is Useless

In the Emergency Department, that is, at least.

A few years ago, there were several major studies evaluating the safety of a CT coronary angiogram-based study for the evaluation of chest pain in the Emergency Department. These studies consistently found the CTCA is sensitive for coronary artery plaque – and, thus, patients with troponin-negative chest pain syndromes and clear coronary arteries could be discharged from the ED.

The constant challenge, however, has been specificity – not only with respect to whether the CT can accurately detect stenoses, but the clinical relevance of the stenoses. Non-obstructive, moderate, and high-grade stenoses detected on CTCA all trigger further evaluation – either non-invasive or invasive, with subsequent interventions of uncertain clinical value.

This small propensity-matched study from a cohort of 25,251 patients undergoing CTCA picked out 234 pairs of patients, matching those who had an acute coronary syndrome during follow-up with those who did not. And, yes, those with ≥50% or ≥70% stenosis were more likely to suffer an ACS, but not my much. The vast majority – 62% – of those with an ACS in follow-up had non-obstructive coronary disease. Indeed, just over half of patients with an ACS even had their culprit lesion identified on the initial CTCA. The degree of stenosis was mildly predictive of future ACS, but plaque burden between those who suffered an event and those who did not was similar. The most predictive feature, however, was composition of non-obstructive plaque, including fibrofatty features and necrotic core.

This is why CTCA is unhelpful in the Emergency Department. It does, yes, accurately detect patients without coronary disease – but this target “low-risk” population already has such a low pretest likelihood of poor outcome the added value is nil. Then, the “true positives” from these studies – stenoses and interventions – are not equivalent to ACS prevented.

Friends don’t let friends do CTCA in the ED – it doesn’t add value or prevent adverse outcomes.

“Coronary Atherosclerotic Precursors of Acute Coronary Syndromes”
http://www.onlinejacc.org/content/71/22/2511

Don’t Give NEXUS II Much Thought For Kids

Into the the world of PECARN, CHALICE, and CATCH, we add NEXUS II. Why? Good question.

This is a planned secondary analysis of the NEXUS Head CT decision instrument among enrolled patients less than 18 years of age. Like most decision instruments, this rule classifies patients into “high risk” or “low risk”, with “low risk” being free of any mandated imaging. Their rule, which I will not recount, was tested in 1,018 blunt head trauma patients, and their rule picked up all 27 of those who required neurosurgical intervention. Unfortunately, it also only classified 330 patients as “low-risk” – for an abysmal 33% specificity.

The authors state it may yet have value, despite this poor specificity, as a one-way decision rule. Unfortunately, one-way decision rules are fraught with peril, as the inability to classify a patient as “low risk” is difficult to ignore.  This leads clinicians to ultimately use the one-way instrument as a two-way, despite the bleak positive predictive value. This rule also missed one of 49 patients with “significant intracranial injuries”, meaning it is reasonable to expect it may not actually be 100% sensitive.  Considering clinical judgement is vastly superior to this product, and there are enough alternative options, it is reasonable not to give this product another thought.

“Validation of the Pediatric NEXUS II Head CT Decision Instrument for Selective Imaging of Pediatric Patients with Blunt Head Trauma”
https://www.ncbi.nlm.nih.gov/pubmed/29665151

A Validation of YEARS?

A couple of years (ha ha!) ago, the results of the YEARS study were unveiled, a culmination of various ideas towards incorporating pretest probability into the use of the D-dimer for “ruling out” pulmonary embolism. The ideas were not unique to this research group in the Netherlands, but the implementation was – and many awaited some independent confirmation of their results.

This is a step towards that independent confirmation, a multi-center observational evaluation of the YEARS protocol. In this study, these authors collected the data necessary to determine the workup and outcomes via YEARS, but clinical practice was left unchanged. This is of particular importance because, unfortunately, this leaves a glaring hole in their enrollment criteria – YEARS did not have any carve outs for patients in whom D-dimer was not part of the evaluation for PE, whereas in these EDs, patients with high pretest likelihood typically went straight to imaging without D-dimers.

Overall, then, in these 17 Emergency Departments, 1,789 low- and intermediate-risk patients were evaluated for PE, with an overall prevalence of 4%. There were 7 (0.3%) patients determined to have potentially missed PE by three month chart review and follow-up phone calls, with a default assignment of “no PE” for the handful lost to follow-up. In their sample, usual practice led 45% undergoing CTPA and 9% undergoing V/Q. With YEARS criteria implemented, the imaging rate would have been 33% – at the cost of an additional 6 missed PEs. This miss rate, however, is still below the threshold of testing equipoise, balancing the morbidity of anticoagulation versus the morbidity of missing PE.

This is a helpful piece of evidence in support of YEARS, even if it leaves major questions unanswered regarding its use in high-pretest probability patients, to say the least. The imaging reduction estimation, as well, should be considered to be overstated by an observational trial, as the adoption of apparently aggressive new protocols would likely be slower than this theoretical maximum.

One small step forward.

“Multicenter Evaluation of the YEARS Criteria in Emergency Department Patients Evaluated for Pulmonary Embolism”
https://www.ncbi.nlm.nih.gov/pubmed/29603819

Nothing Excludes Pulmonary Embolism?

It’s hard to find a diagnostic pathway with greater variation than that of pulmonary embolism. On one hand, you have the YEARS protocol, in which D-dimer is the definitive gatekeeper without carve-outs or exclusions. On the other hand, you have an article like this – saying even CT pulmonary angiograms are inadequate to rule-out PE.

These authors re-analyzed the data from a previous prospective study enrolling 7,940 patients across 12 Emergency Departments. In this analysis, these authors focused in on the 257 patients for whom a “High risk” Wells score was assigned. Of these, 201 underwent reliable CTPA, 71 of which were read as positive and 130 of which were read as negative. Within 45 days, using chart review and telephone follow-up, the authors determined 16 of these patients were ultimately diagnosed with PE. They conclude the CTPA missed these diagnoses at the index visit, and should not be considered adequate for rule-out in a high-risk patient. They go on to cite the inadequate sensitivity of CTPA as demonstrated in PIOPED-II as justification for their stance.

Unfortunately, there’s not nearly enough information presented here to fully evaluate their findings. The authors are attempting to refute the utility of CTPA as a reliable mechanism for rule-out, but, despite such a small sample, no individual scan follow-up was attempted to overread the initial CT. Then, patients with high Wells scores are obviously at high risk for VTE; it almost certainly reasonable some of these downstream PE are independent events, a possibility towards which the authors are fairly dismissive. The authors report many of these patients were positive for DVT, making a subsequent PE as an independent event even more likely. The PIOPED-II study, landmark or not, was conducted in a comparatively medieval era of CTPA, and those sensitivity findings should not impact current data interpretation. Finally, the CTPA is famous not so much for false negatives as it is false positives. The authors do not account for the possibility some of these downstream diagnoses are false positives, and no characterization of the subsequent diagnoses are given – an important consideration in this era of over-diagnosis of subsegmental PE of uncertain clinical significance.

I certainly do not believe these data should change practice or our opinion of the CTPA as a reliable rule-out for a clinically important PE. More robust, prospective study is necessary to confirm the veracity of their conclusion of these false negatives.

“Ruling out Pulmonary Embolism in Patients with High Pretest Probability”

https://escholarship.org/uc/item/74h4h8qb

How Many 6-to-24 Hour Stroke Patients Are Eligible?

So, DAWN and DEFUSE-3 show it is reasonable to use tissue-based criteria to guide intervention, rather than the quaint, but anachronistic, concept of “time is brain.” However, in expanding this window, what is the yield of screening? How many CT cerebral angiograms with specialized perfusion imaging will need to be performed to identify a patient for intervention?

This single-center report 2014 through 2017 at a DAWN trial-participating center found:

  • 2,667 patients with acute ischemic stroke.
  • 792 arrived between 6 and 24 hours of last known to be normal.
  • 298 of those were NIHSS ≥10.
  • 155 of those had proximal anterior large vessel occlusion.
  • 45 of those were non-disabled at baseline and met clinical and imaging mismatch criteria.

The authors also did an analysis for DEFUSE-3 eligibility, and ended up with similar numbers, although there were 15 DAWN-eligible patients who did not meet DEFUSE-3 criteria and 28 DEFUSE-3 patients who did not meet DAWN criteria, so there’s some fuzziness at the bottom of the pyramid, in addition to the limitations of their retrospective review.

So, effectively, a little more than a third of patients presenting between 6 and 24 hours probably meet criteria for screening for large-vessel occlusion, with about half of those identifying an occlusion, and then another third of those having imaging findings with sufficient viable tissue for intervention.

There are almost certainly opportunities to use clinical evaluation – not just a NIHSS cut-off – to improve yield, but there will inevitably be a balance between sensitivity and specificity with respect to resource utilization.

“Eligibility for Endovascular Trial Enrollment in the 6- to 24-Hour Time Window”
http://stroke.ahajournals.org/content/early/2018/03/15/STROKEAHA.117.020273