Chillin’ Children After OHCA

Once upon a time, many adults suffering an out-of-hospital cardiac arrest received therapeutic hypothermia with a target temperature of 33°C.  Then, along came the Targeted Temeperature Managment trial – in which 36°C seemed to be just as good as 33°C.  Now, just to throw another confounder in the mix, we have a trial comparing 33°C to “therapeutic normothermia” – 36.8°C – and we’re doing it in children to address concerns regarding generalizability from adults.

Very detailed summaries of the numbers, methods, and enrollment can be found on other #FOAMed sites – particularly St. Emlyns and ALiEM.  But, the high points:

  • Many – 1,355 – were screened, but ultimately only 260 were randomized and included in their primary analysis.
  • Adherence to temperature management protocols was good or adequate in ~90% of cases.
  • Hypothermia was implemented for 48 hours, followed by normothermia up to 120 hours total to match the normothermia group.
  • In contrast to adults, the great majority (72%) of this pediatric cohort suffered a respiratory arrest.

The outcome: no statistical difference, with 20% of the hypothermia group alive and functional at 1 year, compared with 12% of the normothermia group, a p-value of 0.14.  Regarding safety, arrhythmias and culture-proven infections favored the normothermia group, 1% vs. 5%, and 39% vs. 46%, but these also did not reach statistical significance.  Finally, both 28-day and 1-year mortality favored hypothermia, with an absolute difference of ~10% in each, but this was not statistically significant, either.

I will let the authors speak for me here:

“One important potential limitation of the trial is that, on the basis of the observed confidence limits for treatment differences, a potentially important clinical benefit cannot be ruled out despite the lack of a significant difference in the primary outcome measure. A larger trial might have detected or rejected a smaller intervention effect. Indeed, there was a significant difference in survival time with therapeutic hypothermia, although this was a secondary outcome measure.”

The relative likelihood of benefit for hypothermia in this trial was 1.54, with a 95% CI of 0.86 to 2.76.  Now, this result crosses 1, and therefore requires interpretation in two contexts.  The first is the normal distribution:

In which we visualize the frequency of potential outcomes, and the important realization the more frequent “true” outcome is most likely to occur near the center of the 95% CI range.

And, the more important context:

In which we interpret these data in the context of prior results, generalized from other settings.  In this case, our prevailing opinion is one in which we suspect hypothermia – with much uncertainty regarding the details – is beneficial.  As you can see, the effects of even “statistically significant” findings have only limited practical impact on the “good bet” or the “long shot”.  Hence, the results of this study – which simply barely fail to reject the null hypothesis – do not hardly move the needle against the prevailing opinion.

I tend to side with the authors of this “negative” study: it is mostly likely underpowered to detect the expected benefit, and it is still reasonable to cool children following OHCA.  There are many questions that remain regarding the temperature, duration, and other details – not limited only to children – but it would be erroneous to say this trial refutes the practice of hypothermia in children.

“Therapeutic Hypothermia after Out-of-Hospital Cardiac Arrest in Children”
http://www.nejm.org/doi/full/10.1056/NEJMoa1411480

Where’s the Beef With TXA?

CRASH-2 was a massive, international undertaking, testing the utility of tranexamic acid to improve outcomes in bleeding trauma patients.  When given with 3 hours, there were significant reductions in mortality due to bleeding – and the current push for its widespread use was born.

However, this study, and others like it, is not seeing the same magnitude of success as described in CRASH-2.  This single-center, retrospective evaluation of trauma patients reviewed the mortality benefit associated with implementation of a thromboelastography-based TXA protocol.  In 2011, this institution introduced a TEG-based TXA threshold of estimated percent lysis at 30 minutes of >3.0%, and these authors reviewed all cases of trauma patients between 2009 and 2013 meeting that threshold and eligible for treatment within 3 hours.

These authors identified a cohort of 98 patients who met criteria and received TXA, and compared them with a cohort of 934 patients who met criteria and did not.  In-hospital mortality in the cohort receiving TXA was double those who did not (40% vs. 17%), and this disadvantage persisted despite adjustment for age, gender, mechanism, ISS, hypotension, and base excess.  TXA usage was also not associated with resolution of hyperfibrinolysis, as measured by follow-up TEG, but neither was it associated with an increase in thromboembolic events.

Unfortunately, this retrospective evaluation is biased by confounding and unmeasured selection imbalances.  It is, however, not the only study questioning the value of TXA in the setting of routinely well-resuscitated, modern trauma evaluation.  Nothing in this small review provides compelling evidence regarding cessation or tailoring of TXA therapy in bleeding trauma patients, but it does support its continued evaluation for its role in organized, modern trauma settings.

“The impact of tranexamic acid on mortality in injured patients with hyperfibrinolysis”

Still Not Choosing Wisely in Trauma Imaging

We can all agree the advent of CT has improved our diagnostic capabilities, particularly in multi-system trauma.  Few would challenge an assumption that outcomes are positively impacted by timely, accurate identification of clinically important pathology.

Unfortunately, the pendulum has swung so far in favor of CT in trauma, any intelligent reliance on clinical exam skills has been deprecated to obsolete.  As such, the expected fallout includes increases in costs, radiation, and length-of-stay as the zero-miss culture creeps from multi-system trauma into the lightly injured.  This has become such an issue the American College of Surgeons devoted one of five slots in their first Choosing Wisely Guidelines to reducing the use of the trauma “pan-scan”.

Hopefully, the culture change will happen none-to-soon, as this NHAMCS data review indicates – showing steady increases in CT use for both head and body over the 2007 to 2010 review period.  Head CT increased from 9.6% to 11.6% of all injury-related encounters, while body CT increased from 5.5% to 8.1% – without any corresponding increase in positive findings.  Yield for severe injury dropped from 4.9% to 3.4% on Head CT, along with a drop for body CT from 6.4% to 3.3%.

This is the NHAMCS probabilistic sample, of course, and it’s simply a coarse observational cohort without detailed clinical factors.  However, I think the likelihood these observations accurately reflect reality is rather high.

Choose more wisely, please.

“Trends in Advanced Computed Tomography Use for Injured Patients in United States Emergency Departments: 2007–2010”
http://www.ncbi.nlm.nih.gov/pubmed/25996245

Oxygen: Friend or Foe?

I’m a huge fan of oxygen.  I breathe oxygen nearly every day, and without it, I would literally, moreso than figuratively, die.

But, oxygen is a highly reactive molecule with many adverse effects in the human body.  Recognition of such seems to be in direct contrast to the otherwise reasonable hypothesis of increased oxygenation providing benefit in ischemic disease states.  The most recognizable of these is acute myocardial infarction, where oxygen is enshrined in the classical (and outdated) MONA mnemonic.

This is the AVOID trial, randomizing patients in the field with prehospital diagnosis of STEMI to either 8L/min inhaled oxygen, or oxygen only as needed to maintain saturations >94%.  All patients received aspirin from paramedics en route to the receiving facility, with further care as per local standards and protocols.  The primary outcome was infarct size, as measured by peak troponin and creatine kinase levels.

Paramedics screened 836, randomized 638, an additional 50 were protocol non-compliant, and then 118 were declared not to be STEMI upon arrival at the receiving facility.  The remaining 470 underwent angiography, and the final cohort for analysis was the 441 for whom STEMI was ultimately confirmed.  Groups were generally similar between interventions, although there was an excess of 8 patients with LAD lesions in the oxygen arm and of 10 patients with circumflex lesions in the no-oxygen arm.  There were 11 excess single-vessel patients in the oxygen arm and 17 excess multi-vessel disease patients in the no-oxygen arm.

The answer?  Oxygen is probably bad.  There was no statistically significant difference in mean peak troponin values, favoring the no-oxygen having a p-value of 0.18.  The mean peak CK difference did, however, reach significance, with a p-value of 0.01.  In the 127 patients for whom follow-up MRI imaging was available, measures of infarct size all favored the no-oxygen group, with p-values ranging between 0.04 and 0.08.  However, clinical outcomes were all over the map.  The no-oxygen arm had higher in-hospital mortality, but the oxygen arm had higher rates of recurrent myocardial infarction.  Long-term, six-month outcomes were likewise similar, with trivial clinical differences.

So, oxygen application during routine pre-hospital transport for chest pain is certainly useless and wasteful – and most likely at least a little bit harmful.

“Air Versus Oxygen in ST-Segment Elevation Myocardial Infarction”
http://circ.ahajournals.org/content/early/2015/05/22/CIRCULATIONAHA.114.014494.abstract

Your Bouncebacks Are Not Alone

“Remember that patient you had yesterday?” is infrequently a favorable start to a conversation.  Emergency Department bouncebacks are frequently tracked metric, ostensibly for self-reflection, but also as a proxy for care quality and mismanagement.

This is a 6 state review of 2 to 5 years of data linked between State Emergency Department Databases and State Inpatient Databases, evaluating Emergency Department recidivism up to 30 days.  The authors also linked this data to healthcare cost data, but highest quality cut of meat here is the detail on bouncebacks.  Based on 53,530,443 Emergency Department visits, these authors found the overall 3-day revisit rate was 8.2%, and the 30-day revisit rate was 19.9%.  Approximately 2/3rds of revisits were to the same Emergency Department, with the remainder choosing a different ED.

These numbers, I think, are much higher than most would expect – and provide at least a small amount of solace if you feel as though it seems there’s always a previous patient of yours checking back into the ED.  The authors break down several interesting details regarding the types of revisits:

  • Skin and soft-tissue infections resulted in 23.1% 3-day revisit rates, with 12.9% admission on revisit.
  • Abdominal pain was the second-most frequent revisit, at 9.7%, associated with 29.9% admission on revisit.
  • Patients aged 18-44 were more likely to visit a different ED for the second visit, while patients aged 65 and above were the most likely to be admitted on revisit.
  • Patient with back pain had the highest revisit rate to a different ED within 3 days, 7.8%, with 41% of those visiting a different ED.

Simply at face value, these additional visits are expensive and resource-intensive – particularly if there’s not an effective local electronic information exchange preventing duplication of testing.  There is also clearly ample opportunity to develop targeted interventions for certain groups of patients to potentially provide follow-up care in a lower-cost setting.

“Revisit Rates and Associated Costs After an Emergency Department Encounter”
http://www.ncbi.nlm.nih.gov/pubmed/26030633

Why Do We Still Admit Chest Pain?

If you worked a shift today, you had a patient with chest pain.  As these authors cite in their introduction, visits for chest pain comprise 1 in 20 presentations to Emergency Departments – and the evaluation of such patients costs more than the annual GDP of Malta.  As our hospitalist colleagues lament, a massive subset of inpatient evaluations for chest pain are invariably negative – or, even worse, generate false positives and other iatrogenic harms.

This study is an retrospective evaluation of an observational registry of chest pain presentations to three Ohio Emergency Departments.  The authors perform a search of five years worth of data, and generate a cohort consisting of patients who received at least two consecutive negative troponins initiated in the Emergency Department.  The primary outcome was in-hospital life-threatening arrhythmia, STEMI, cardiac arrest, or death.

In this database of 45,416 patients, 11,230 met their inclusion criteria.  Independent, hypothesis-blinded abstractors reviewed a subset of “possible” primary outcomes based on electronic data, and manually abstracted those identified.  From this manual review, there were 20 (0.18%) patients for whom a critical outcome was identified.  The authors reviewed each specific case and tried to identify specific risks for adverse outcome – and, if patients with abnormal vital signs, left bundle branch block, pacemaker rhythm, or signs of EKG ischemia were further excluded, the incidence of critical outcomes drops to 4 out of 7,266 (0.06%).

The supposed takeaway from this article is that patients who have been ruled out by serial troponin testing have uneventful hospital courses.  Extending this to practice, the theory is we could perhaps generalize this evidence to our 1- or 2-hour rapid-biomarker rule-outs.  These patients would then supposedly have such an acceptable safety profile as could be discharged from the ED with outpatient follow-up to assess the need, or appropriateness, for further provocative or anatomic testing.

These data are not quite strong enough to claim such a strategy as bulletproof.  The risk, I agree, is certainly small – with thousands requiring hospitalization in order to obtain benefit for one patient.  The benefit, however, for the patients in this study is not the soft MACE outcome described in other studies – these are hard endpoints of folks who would likely be dead if not observed in the hospital.

While I expect outpatient evaluation of substantial numbers of chest pain patients to be the new culture in Emergency Medicine in the future – and as much as I would like to purchase Malta for ACEP next year – this isn’t zero-miss.  These data support development of appropriate outpatient strategies – but not wholesale practice revision based solely on this data set.

Addendum: Louise Cullen makes a few excellent points on social media peer review I’ll paraphrase here: 1) The endpoints measured here are not the only important patient-oriented outcomes.  There are a small number of initially troponin-negative acute coronary syndromes that may be missed here. 2) There are patients for whom hospitalization and urgent evaluation has value due to medical interventions initiated in the hospital.  An aggressive discharge strategy cannot be based on a catch-and-release foundation without tightly integrated follow-up.

“Risk for Clinically Relevant Adverse Cardiac Events in Patients With Chest Pain at Hospital Admission”
http://archinte.jamanetwork.com/article.aspx?articleid=2294235

The Very Young Pediatric C-Spine Rarely Needs Radiologic Clearance

It is usually reasonable to exercise an abundance of caution with trauma patients suspected of having cervical spine injuries.  However, an abundance of caution sometimes means an abundance of radiation – and the costs and harms associated with such testing can be immense, regardless of technical difficulty in a young pediatric population.

This is a retrospective evaluation of 2,972 trauma patients aged less than 5 years, reviewing specifically the overall incidence of diagnosed cervical spine injury.  In this 12 year cohort, a grand total of 22 had confirmed CSI.  Most importantly, however, nearly all cases of CSI were associated with other serious injuries – a cohort with a median ISS of 33.  Twelve of 22 arrived intubated, 13 were in extremis, and overall mortality was 50%.  All evaluable patients had either neurologic deficits, severe neck pain, or were unable to range their neck.

These authors do not further describe their cohort for evaluation with regard to developing a predictive instrument for cervical spine injury, but these data do support a very reasonable conclusion regarding the rarity of pediatric injuries – and the near impossibility of isolated cervical spine injuries.  I tend to agree with the authors’ stated management strategy for such patients:

“Pediatric patients with abnormal neurologic examination result, decreased mental status, neck pain, or torticollis are evaluated with cervical spine CT; however if the child is asymptomatic defined by a normal neurologic examination result, appropriate mental status, with absence of neck pain or torticollis, our first step is to remove the cervical collar. We examine the patient for cervical tenderness if they are able to communicate and observe the child for normal range of motion of the neck. In preverbal patients, we simply observe neck range of motion with the collar removed. If the child seems to move his or her neck without discomfort and full range of motion, then we do not pursue any further radiologic evaluation.”

“Absence of clinical findings reliably excludes unstable cervical spine injuries in children 5 years or younger”
http://www.ncbi.nlm.nih.gov/pubmed/25909413

DIAS-3 – Desmoteplase Fails in the Extended Time Window

It seems to be the stroke neurologists’ greatest lament – the restricted time windows for tPA, either 3 or 4.5 hours, excluding so many patients from receiving the blessing of thrombolysis.  There have been failed trials in the past in extended time windows, and, even, failed trials in the 3-5h time window.  But, this is desmoteplase, and it is more fibrin specific than alteplase – and this follows up DIAS-2, which seemed to suggest benefit in patients with demonstrated arterial conclusion on vascular imaging.

It is, sadly, negative by the primary outcome of 90-day Rankin score (mRS 0-2), adding another tick mark to the list of failed contemporary trials for systemic thrombolysis.  Safety outcomes, mercifully for the patients involved, were similar, with low rates of neurologic worsening associated with intracranial hemorrhage in each cohort.

The authors, as before, find and focus on a single positive subgroup: patients with ischemic injury volume of less than 25mL on MRI.  There was, interestingly, no positive effect noted for patients whose ischemic injury volume was less than 25mL on CT – and the authors had no specific explanation for the discrepancy.  However, given the recent successful endovascular trials, it is quite reasonable to suggest an imaging-based, tissue-salvage model is more appropriate than the simplistic time-based model suggested by NINDS.  Unfortunately, tissue salvage is dependent upon recanalization – and rates were not significantly different between cohorts, 49% with desmoteplase vs. 42% with placebo.  This is the persistent elephant neurologists fail to acknowledge – that systemic thrombolysis simply rarely works as advertised – greatly diminishing any possible beneficial effect.

The conflict-of-interest statement falls on what probably would have once been considered the extreme side, but now is tragically routine:

The funder was involved in the study design, data collection, data analysis, and data interpretation. Two employees of the funder provided medical writing assistance in the editing of the report. The corresponding author had full access to all study data; all other authors without funder affiliation had access to study data via the corresponding author and authors with funder affiliation had full access to all study data.

Interestingly, review of the ClinicalTrials.gov registration indicates the study was initially planned in 2008 to enroll 320 patients, with an end date in 2010.  In 2010, the planned enrollment was increased to 400, and the study ultimately enrolled 492.  Given the COI involved, it reasonable to suggest the funder was involved in ongoing analysis of the results with the intention of stopping the study at the precise moment a positive outcome – or ultimate futility – was detected.  Despite the best efforts of Jeff Drazen and the NEJM to downplay potential distortions secondary to funding sources, clearly, our vigilance for such likely scientific misconduct should not be diminished.

“Safety and efficacy of desmoteplase given 3–9 h after ischaemic stroke in patients with occlusion or high-grade stenosis in major cerebral arteries (DIAS-3): a double-blind, randomised, placebo-controlled phase 3 trial”
http://www.ncbi.nlm.nih.gov/pubmed/25937443

Where Should You Admit the Elderly with Rib Fractures?

In general, trauma results in disproportionately severe injuries in elderly patients despite similar mechanisms.  A frequent concern, specifically, is the risk of pneumonia or intubation associated with rib fractures – previously demonstrated to increase linear for each fracture in patients aged greater than 65.

However, a dichotomous age cut-off paired with a single variable is an obviously simplistic model.  These authors retrospectively reviewed 400 patients aged greater than 55 years of age hospitalized with injuries including rib fractures.  A regression model was developed to determine predictors of respiratory failure or pulmonary complications.

Six variables shook out of their analysis as significant predictors of subsequent complications:  COPD, low serum albumin, use of an ambulatory assist device, a tube thoracostomy in place, injury severity score, and total number of rib fractures.  Transforming these variables into a scoring system resulted in a predictive instrument with and AUC of 0.82 (0.77 – 0.88), with sensitivities and specificities in the 70%s based on their chosen threshold.

While this performance is suboptimal, the model has obvious face validity – the frail, severely injured, with underlying pulmonary disease are the most likely to deteriorate.  Their model also requires external validation.  However, given that most patients do develop pulmonary complications, such a tool could be reasonably used to reduce the costs and resource utilization associated with prophylactic ICU admissions – as long as you were willing to accept the risk of approximately 1 in 20 patients requiring unexpected escalation in care from the floor.

“A pilot single-institution predictive model to guide rib fracture management in elderly patients”
http://www.ncbi.nlm.nih.gov/pubmed/25909417

Finally, an End to Tamulosin for Renal Colic?

Most urologic professional societies recommend “medical expulsive therapy” for ureterolithiasis, with an expectation of increased stone expulsion, improved time-to-passage, and reduced need for analgesia.

As I’ve covered before – breaking down a pro-tamulosin Cochrane Review – the evidence in support of this practice is junk.  David Newman, Anand Swaminathan, and Salim Rezaie agree.  The last time I posted, I posited there was probably some small benefit to a subgroup of patient with renal colic, but, alas, we would probably never have high-quality evidence.

I was wrong.

This study in The Lancet tested MET by randomizing patients with CT-confirmed ureterolithiasis to three arms – placebo, nifedipine, or tamulosin.  The randomization algorithm balanced the arms between stone size and stone location.  The primary outcome was need for urologic intervention at 4 weeks, with secondary outcomes of patient-reported time to stone passage and pain medication use.

With 1,167 patients randomized – 31 of which were excluded or lost to follow-up – there was no difference in need for urologic intervention between groups: 20% placebo, 19% tamulosin, 20% nifedipine.  Secondary outcomes – measured by follow-up questionnaire – were likewise similar, with no differences detected in the number of pain medication nor days until stone passage.

Now, urologic intervention is a rather imprecise surrogate outcome for evaluating the efficacy of MET for promoting stone passage.  And, only 62% of patients returned the surveys regarding the secondary outcomes of subjective stone passage and analgesic use.  This is high-quality evidence, but hardly infalliable.  The authors also state no subgroup showed benefit – which is not entirely true.  MET was slightly beneficial (86% vs. 82%) for patients with lower ureteral tract stones, with a p-value of 0.099.  Giving into the tyranny of p-values, yes, there’s no benefit – but using the p-value akin to a likelihood ratio, judged against the larger context of other (albeit, low-quality) trials showing benefit, I would not find it unreasonable to contest the totality of these authors’ conclusion.

Regardless, the empiric use of tamulosin has simply been an urban legend taken one step too far.  Short of large stones in the lower urinary tract, the benefit is fleeting at best – and the magnitude of the benefit may be too low to matter.

“Medical expulsive therapy in adults with ureteric colic: a multicentre, randomised, placebo-controlled trial”
http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(15)60933-3/abstract (oa)