If A Tree Falls In A Forest…

After yesterday’s analysis of a not-so-good pan-scan vs selective-scan abstract, it’s time for a little philosophy today. There seem to be two camps in the world of initial diagnostic testing for trauma: selective scanning vs scan everything. I admit that I am one of the former. Yes, the more tests you do, the more things you will find. Some will be red herrings. Some may be true positives, but are they important? Here’s the key question:

“If a tree falls in a forest and no one is around, does it make a sound?”

There is a clinical corollary to this question in the field of trauma:

“If an injury exists but no one diagnoses it, does it make a difference (if there would be no change in treatment)?”

Here’s an example. On occasion, my colleagues want to order diagnostic studies that won’t make any clinical difference, in my opinion. A prime example is getting a chest CT after a simple blunt assault. A plain chest xray is routine, and if injuries are seen or the physical exam points to certain diagnoses, appropriate interventions should be taken. But adding a chest CT does not help. Nothing more than the usual pain management, pulmonary toilet, and an occasional chest tube will be needed, and those can be determined without the CT.

Trauma professionals need to realize that we don’t need to know absolutely every diagnosis that a patient has. Ones that need no treatment are of academic interest only, and can lead to accidental injury if we look for them too hard (radiation exposure, contrast reaction, extravasation into soft tissues to name a few). This is how we get started on the path to “defensive medicine.”

Bottom line: Think hard about every test you order. Consider what you are looking for, what you might find, and if it will change your management in any way. If it could, go ahead. But always consider the benefits versus the potential risks, or what I call the “juice to squeeze ratio.”

References:

  • George Berkeley, A Treatise Concerning the Principles of Human Knowledge, 1734, section 45.
  • paraphrased by William Fossett, Natural States, 1754.

Source: The Trauma Professionals’s Blog

ACS Trauma Abstracts #5: Pan-scan vs Selective CT For Pediatric Patients

In major adult blunt trauma, there are still two factions: those who pan-scan for diagnosis, and those who use CT selectively. The pan-scan proponents argue that they find everything, including things that would have been missed with selective scanning. The selective scanners argue that those things are typically not clinically significant, and radiation exposure is reduced.

Who is right? We’re still not sure. But when it comes to children, most believe that less radiation is always better. The group at USC decided to look at their experience with pan-scan vs selective imaging in blunt pediatric trauma patients, defined as those < 16 years of age. They reviewed their experience over a three year period, excluding those who had low blood pressure (<90). A total of 648 children met these criteria, and an array of variables were analyzed to try to determine “superiority.”

Here are the factoids:

  • 88% of these patients had low injury severity (ISS < 15), 567 patients
  • The low severity group who were selectively scanned had a half-day decrease in length of stay*, a quarter-day decrease in ventilator days, lower morbidity (15% vs 24%)*, and radiation exposure (8 vs 18 mSv)*, with the asterisked variables being “statistically significant”
  • The high severity also showed positive differences in decreased ICU length of stay, ventilator days, morbidity*, and radiation exposure*
  • For both pan- and selective-scanning, additional imaging led to no additional interventions in 95% of cases

Bottom line: Ugh! This is not a good abstract. It shows association, but not causation with anything but the radiation exposure calculations. Yes, if you scan less in the selective arm, there better be less exposure. But the other variables are either not clinically significant, or not defined well (i.e. morbidity).

The authors conclude that selective scanning is the way to go based on this (extremely weak) data. This is why I always recommend that you read the whole paper, not just the abstract, or in this case the whole abstract and not just the conclusion. The data, as presented in this abstract, do not support this at all.

The authors don’t say how many of these patients were very young, and how many of them actually received pan-scans. But any pediatric trauma surgeon would cringe if they read this  article. Although you may be a big believer in pan-scanning, reserve it for adults only until we have some better data. Don’t irradiate kids unnecessarily!

Reference: Selective CT imaging is superior to liberal CT imaging in the hemodynamically normal pediatric blunt trauma patient. JACS 225(4S1):562, 2017.

Source: The Trauma Professionals’s Blog

ACS Trauma Abstracts #4: Timing Of DVT Prophylaxis In Spine Trauma

Spine trauma is one of the high-risk indicators for deep venous thrombosis (DVT). Unfortunately, there is a great deal of variability in the start time for chemical prophylaxis for this injury, especially after the patient has undergone surgery. In part, this is due to a lack of good literature and guidelines, and in part due to the preferences of the spine surgeons who operate  on these patients.

A group at the University of Arizona in Tucson performed a large database review (looks like National Trauma Databank, although they don’t say in the abstract) looking at “early” vs “late” administration of prophylaxis after surgery in these patients. The spine injury was the predominant one, with all other systems having an abbreviated injury score (AIS) < 3. They matched two years worth of patients for demographics, initial vitals, type of operative intervention, and type of heparin to assess the impact of prophylaxis timing.

Here are the factoids:

  • Nearly 40,000 patient records were reviewed, and over 9,500 met the spine injury criteria with operation and prophylaxis. A total of 3,556 could be matched for analysis.
  • These patients were split in half for matching, late (>48 hrs) versus early (<48 hrs)
  • DVT rate was significantly lowe in the early prophylaxis group (2% vs 11%)
  • PE rate and mortality were the same between groups
  • Return to OR and blood transfusion rates were identical (1% and 1-2 units)

Bottom line: Once again, we see that “early” prophylaxis for DVT is probably desirable and mostly harmless, even after a spine operation. Many surgeons still have an irrational fear of giving heparin products in patients who have some risk of bleeding. The body of literature that supports early use just keeps growing. One observation, though: as in most other studies, pretty much whatever we do for DVT has a negligible impact on PE and mortality. We can only treat the clots, but not their major aftermath.

Reference:  Optimal timing of initiation of thromboprophylaxis in spinal trauma after operative intervention: – propensity-matched analysis. JACS 225(4S1):S59-S69, 2017.

Source: The Trauma Professionals’s Blog

ACS Trauma Abstracts #3: Using Mechanism Criteria To Activate The Trauma Team

Most US trauma centers have two tiers of trauma activation. The higher tier is typically called for physiologic derangements like hypotension, tachypnea, or decreased mental status. This triggers arrival of the full trauma team for rapid assessment and management.

The second tier is reserved for patients who may be less seriously injured and usually results in a reduced team. And depending on how good the activation criteria for this tier are, many patients eventually turn out to have no serious injuries and are discharged from the emergency department. This is the purest form of overtriage, and if it occurs frequently, can wear down your trauma team and waste resources.

Criteria for the second tier trauma activation may include mechanism of injury criteria such as ejection, pedestrian struck, intrusion into the passenger compartment, death at the scene, and other similar criteria. They sound like good criteria, but how helpful are they, really? The group at Baylor University Medical Center in Dallas performed a retrospective review of their trauma activations over a one and a half year period to test the efficacy of some of these criteria. They had recently added some mechanism-based criteria to their second tier activations.

Here are the factoids:

  • During the study period, they had 1325 second tier activations, and 603 were based on mechanism criteria
  • The mean injury severity score of mechanism-based criteria was only 5, versus 10 for anatomic criteria (significant)
  • A whopping 37% of mechanistic criteria patients were discharged home from the ED, versus only 10% for other criteria (also significant)
  • Second tier activations for physician discretion were just as good as non-mechanism criteria, with an ISS of 10 and 13% discharged home
  • Looking at specific criteria, compartment intrusion, ejection, and death in the compartment appeared to be the major overtriage offenders, with an ISS of 5 and 40% discharge rate
  • Incidentally, penetrating injury proximal to knee or elbow had very high overtriage rates, with an ISS of 1 and discharge rate of 48%

Bottom line: Trauma centers are encouraged to review their trauma triage criteria periodically, especially when overtriage rates are high. This center is presenting a nice paper that shows the benefit of doing this. They identified four mechanistic criteria that do not appear to be any better than just relying on physician discretion. What they are not saying is that it is probably better to rely on physiologic and anatomic criteria, as well as physician discretion, to determine which level of trauma activation to trigger.

And incidentally, the new ACS highest-level criterion of gunshot proximal to knee or elbow may not be everything its cracked up to be. It’s difficult to say for sure because stabs and gunshots were not separated out in this abstract, and the number they encountered was not specified. But it certainly suggests this criterion needs some fine-tuning as well.

Reference: Intrusion, ejection, and death in the compartment: mechanism-based trauma activation criteria fail to identify seriously injured patients. JACS 225(4S1):S56, 2017.

Source: The Trauma Professionals’s Blog

Anticoagulants And The Elderly: Are They Being Appropriately Treated?

About 2.3 million people, or a bit less than 1% of the US population, have atrial fibrillation. This condition is commonly managed with anticoagulants to reduce the risk of stroke. Unfortunately, the elderly represent a large subset of those with a-fib. And the older we get, the more likely we are to fall. About half of those over 80 will fall once a year.

Are all of these elderly patients being treated with anticoagulants appropriately? Several scoring systems have been developed that allow us to predict the likelihood of ischemic stroke. Looking at it another way, they allow us to judge the appropriateness of using an anticoagulant to prevent such an event.

The original CHADS2 score was developed using retrospective Medicare data in the US. The newer CHA2DS2-VASC score used prospective data from multiple countries. However, the accuracy is about the same as the original CHADS2 score. But because the newer system has three more variables, it adds a few more people to the high-risk group who should receive an anticoagulant.

The higher the CHA2DS2-VASC score, the more likely one is to have an ischemic stroke. The threshold to justify anticoagulation seems to vary a bit, with some saying >1 and others going with >2. Here’s a chart that shows how the stroke risk increases.


Stroke risk per year with CHA2DS2-VASC score

Whereas CHA2DS2-VASC predicts the risk of clotting (ischemic stroke), the HAS-BLED score looks at the risk of bleeding. It includes clinical conditions, labile INR, and concomitant use of NSAIDs, aspirin or alcohol, but not a history of falls.

Proper management of atrial fibrillation in the elderly must carefully balance both of these risks to reduce potential harm as much as possible. A HAS-BLED score of >3 indicates a need to clinically review the risk-benefit ratio of anticoagulation. It does not provide an absolute threshold to stop it.

A group at Henry Ford Hospital in Detroit, a Level I trauma center, retrospectively reviewed their experience with patients who fell while taking an anticoagulant for atrial fibrillation. They calculated CHA2DS2-VASC and HAS-BLED for each and evaluated the appropriateness of their anticoagulation regimen.

Here are the factoids:

  • A total of 242 patients were reviewed, and the average age was 78
  • The average CHA2DS2-VASC score was 5, and the average HAS-BLED was 3
  • Only 1.6% were considered to be receiving an anticoagulant inappropriately (CHA2DS2-VASC 0 or 1)
  • Nearly 9% of patients were dead 30 days after the fall

Bottom line: The authors found that their population was appropriately anticoagulated. But they also noted that the morbidity and mortality risk was high, and was independent of age and comorbidities.

There are tools available to help us judge whether an elderly patient should be taking an anticoagulant for atrial fibrillation. The tool for predicting bleeding risk, however, is not as good for trauma patients. It ignores the added risk from falling, which is very common in the elderly.

Every patient admitted to the trauma service after a fall should have a critical assessment of their need for anticoagulation. The specific drug they are taking (reversible vs irreversible) should also be examined. If there is any question regarding appropriateness, the primary care provider should be contacted personally to discuss and modify their drug regimen. Don’t just rely on them reading the hospital discharge summary. Falls can be and are frequently fatal, just not immediately. Inappropriate use of anticoagulants can certainly contribute to this problem, so do your part to reduce that risk.

Related links and posts:

Reference: Falls, anticoagulation, and the elderly: are we inappropriately treating atrial fibrillation in this high-risk population? JACS 225(4S1):S53-S54, 2017.

Source: The Trauma Professionals’s Blog

ACS Trauma Abstracts #1: REBOA! (And CT???)

This paper is from the group at ShockTrauma in Baltimore, who are really pushing the envelope of REBOA. We always worry about distal ischemia after balloon inflation, because the ischemia produced can be detrimental to the gut and lower extremities. This group was curious about what the flow patterns looked like with  inflation of the balloon. So in select cases, they obtained CT scans with contrast in patients while the balloon was fully inflated (!!).

They reviewed their experience over a four year period, looking at patients receiving a CT scan with the REBOA balloon partially or fully inflated.

Here are the factoids:

  • Nine patients were included. This makes sense because unstable patients should not go to CT scan, so this should be a very limited group.
  • Mean injury severity score (ISS) was 48, which makes sense. These patients are hurt bad!
  • Four patients had supraceliac REBOA (aortic zone I) and five had infrarenal (zone III)
  • Contrast was seen below the REBOA balloon in all patients, and was seen distal to the insertion site in half
  • Collateral flow around the balloon was identified in all patients

Bottom line: The authors found that REBOA decreased blood flow to the distal aorta, but certainly did not stop it. Collateral flow is underestimated, and probably provides a protective effect for the viscera and other structures while inflated. This is good news for REBOA proponents, because it suggests that placement may not cause as much risk from ischemia as originally thought.

But why oh why did they have to go to  CT in the first place?

Reference: Assessment of blood flow patterns distal to aortic occlusion (AO) using computed tomography in patients with resuscitative endovascular balloon occlusion of the aorta. JACS 225(4S1):S50, 2017.

Source: The Trauma Professionals’s Blog

Best Of: Abstracts From The American College of Surgeons

The annual Clinical Congress for the American College of Surgeons will be held at the end of the month. A large number of research papers are presented there, and a smaller subset are related to trauma and critical care.

I’m going to spend the next week and a half or so reviewing many of the ones I believe are the most promising. Keep in mind that these are works in progress, and that many will never see the light of print. I’ll take them apart, trying to see how good each one is. Some are very good, some are hopelessly flawed.

I hope you enjoy the analysis! Let’s see if we learn something new along the way.

Source: The Trauma Professionals’s Blog

EMS: Scoop and Run or Stay and Play for Trauma Care? The Final Answer?

I’ve been discussing the little research there is on stay and play vs scoop and run. And now, hot off the press, we have a paper about the ultimate version of scoop and run, the “drive-by ambulance.” This one looks at outcomes in patients who are dropped at the emergency department by private vehicle. This is the most basic form of prehospital care, with no interventions, just transportation. This type of transport is used by parents with their injured children, police who act as very basic first responders in some cities, and on occasion, gang members.

A multi-institutional group of authors used data in the National Trauma Databank to try to answer this question. They used three years worth of data, comparing outcomes from patients with ground EMS vs private transport who were treated at Level I and Level II trauma centers. Only gunshots and stabs were included, and all patients were 16 or older.  The  authors were focused on only one thing: mortality. This included death in the ED, and a model was developed to adjust risk based on vital signs, injury severity score (ISS), race, and insurance status. Just over 100,000 were included in he final analysis.

Here are the factoids:

  • Black and hispanic patients were more frequently transported by private vehicle, but only by about 3%
  • White patients were more frequently taken by ground EMS, by about 6%
  • Stabs were more likely to be brought by private car than gunshots, 56% vs 44%
  • Mean ISS was significantly higher for those transported by ground EMS (10 vs 5.5)
  • Unadjusted mortality was lower overall for private vehicle, 2% vs 12%
  • The chance of dying in the ED was also decreased in the private transport patients, from 7% to 1%
  • Mortality from both gunshots and stabs were also significantly lower (5% vs 9%, and 0.2% vs 3% respectively)
  • Once adjusted for risk, the lower mortality with private transport remained, with a 60% risk reduction of death. This persisted in the gunshot and stab cohorts as well.

Bottom line: Wow! Although this study has the usual limitations of using a large external database, it was very well designed to compensate for that. And the degree of improvement in survival is surprising. What this study can’t tell us is why. Certainly, some patients benefit from a little extra time to give fluid or blood, intubate, or provide some other treatments. But the patient with penetrating frequently does not need this, they just need definitive control of hemorrhage. The authors even go so far as to suggest that at-risk populations receive education on “scoop and run” if people they know suffer penetrating injury.

The next step is to tease out which components of the stay and play paradigm are the most valuable, and which contribute to the increased mortality in penetrating injury patients.

Reference: Association of Prehospital Mode of Transport With Mortality in Penetrating Trauma A Trauma System–Level Assessment of Private Vehicle Transportation vs Ground Emergency Medical Services. JAMA Surgery, Published online September 20, 2017.

Source: The Trauma Professionals’s Blog

EMS: Scoop and Run or Stay and Play for Trauma Care? Part 3

Scoop and run or stay and play. Is one better that the other? Over my last two posts, I reviewed a couple of papers that were older (6-7 years) and had smaller patient groups. Now let’s look at a more recent one with a larger experience using a state trauma registry.

This one is from the Universities of Pittsburgh and Rochester, and used the Pennsylvania state trauma registry for study material. The authors wanted to really slice and dice the data, postulating that previous studies were not granular enough, such that significant trends could not be seen due to lumping all prehospital time together. They divided prehospital time into three components: response time, scene time, and transport time. To some degree, the first and third components are outside of the prehospital providers’ control.

The records for over 164,000 patients were analyzed. These only included those for patients transported from the scene by EMS, and excluded burns. The prehospital time (PH time) was divided into the three components above. A component was determined to be prolonged if it contributed > 50% of the total PH time.

Here are the factoids:

  • Half of the patients had a prolonged PH time interval (52%)
  • Response time was prolonged in only 2%, scene time was prolonged in 19%, and transport time was longer in 31%
  • Mortality was 21% higher in those with a prolonged scene time component
  • There was no mortality difference in patients with no prolonged time components, or those with prolonged response or transport times
  • These patterns held for both blunt and penetrating injury
  • Extrication and intubation were common reasons for prolonged scene time. Extrication added an average of 4.5 minutes, and intubation 6.5 minutes.
  • Mortality was increased with prehospital intubation, but this effect lessened in severe TBI
  • Increasing experience with extrication and intubation appeared to decrease the mortality from the increased scene time they caused

Bottom line: This paper suggests that the dichotomy of “scoop and run” vs “stay and play” may be too crude, and that a more nuanced approach should be considered. In plain English, the optimal management lies somewhere in between these polar opposites. Actual on scene time appears to be the key interval. EMS providers need to be aware of scene time relative to response and transport times. Patients with specific injury patterns that benefit from short scene times (hypotension, flail, penetrating injury) can quickly be identified and care expedited. Increased scene time due extrication cannot be avoided, but prehospital intubation needs to be considered carefully due to the potential to increase mortality in select patients. 

Reference: Not all prehospital time is equal: Influence of scene time on mortality. J Trauma 81(1):93-100, 2016.

Source: The Trauma Professionals’s Blog

EMS: Scoop and Run or Stay and Play for Trauma Care? Part 2

Yesterday, we looked at an older study that kind of examined the scoop and run vs stay and play debate.  Let’s move forward in time a little bit, and evaluate the two options in a penetrating trauma model.

This one is from the anesthesia and intensive care departments at the university hospital in Copenhagen. The authors prospectively captured information on 462 penetrating trauma victims, then looked up their 30 day survival status in a national administrative database.

Here are the factoids:

  • Only 95% of patient records (446) were available for 30 day review (better that in the US!)
  • Of those, 40 were dead (9%)
  • Using raw statistics, there seemed to be a significant increase in mortality if the prehospital crew was on scene more than 20 minutes
  • However, when corrected for age, sex, injury pattern, etc. there was no significant difference in survival for short vs longer scene stays
  • Multivariate analysis identified the number of procedures performed at the scene as a significant predictor of mortality, regardless of time

Bottom line: We still can’t seem to show a difference in patients who are tossed in the back of the squad and driven vs those who have IVs, immobilization, and other things done to begin resuscitation and increase safety prior to transport! However, the bit about number of procedures is intriguing. Is this just another surrogate for time? Are there unrecognized complications from them that affect survival?

Tomorrow I’ll look at a recent publication from the US that gives us yet another angle on this question.

Reference: On-scene time and outcome after penetrating trauma: an observational study. Emerg Med J 28(9):87-801, 2011.

Source: The Trauma Professionals’s Blog

The Regions Trauma Website