Monday, September 17, 2012

Longer Resuscitation "Saves"

This article made the rounds a couple weeks ago in the news media, probably based on the conclusion from the abstract stating "efforts to systematically increase the duration of resuscitation could improve survival in this high-risk population."

They base this statement off a retrospective review of prospectively gathered standardized data from in-hospital cardiac arrests.  Comparing 31,000 patients with ROSC following an initial episode of cardiac arrest with a cohort of 33,000 who did not have ROSC – the authors found that patients who arrested at hospitals with higher median resuscitation times were more likely to have ROSC.  Initial ROSC was tied to survival to discharge, where hospitals with the shortest median resuscitation time having a 14.5% adjusted survival compared to 16.2% at hospitals with the longest resuscitations.

Now, if you're a glass half-full sort of person, "could improve survival" sounds like an endorsement.  However, when we're conjuring up hypotheses and associations from retrospective data, it's important to re-read every instance of "could" and "might" as "could not" and "might not".  They also performed a horde of patient-related covariates, which gives some scope of the difficulty of weeding out a significant finding from the confounders.  The most glaring difference in their baseline characteristics was the 6% absolute difference in witnessed arrest – which if not accounted for properly could nearly explain the entirety of their outcomes difference.

It's also to consider the unintended consequences of their statement.  What does it mean to continue resuscitation past the point it is judged clinically appropriate?  What sort of potentially well-meaning policies might this entail?  What are the harms to other patients in the facility if nursing and physician resources are increasingly tied up in (mostly) futile resuscitations?  How much additional healthcare costs will result from additional successful ROSC – most of whom are still not neurologically intact survivors?

"Duration of resuscitation efforts and survival after in-hospital cardiac arrest: an observational study"

Friday, September 14, 2012

How Preposterous News Propagates

Every so often – perhaps more frequently, if you're continuously canvassing the literature – there's a rapturous press release regarding a new medical innovation that seems too good to be true.  And, you wonder, how does the lay media get it so wrong?

This study reviewed a consecutive convenience sample of published literature, looking for articles resulting in press releases.  Then, they looked for elements of the article that made it into the press release, as well as the relative accuracy of the release compared with the overall findings of the article.  Essentially, what they found is that press releases were most likely to have "spin" if the conclusion of the article abstract misrepresented the study findings with "spin".

The authors also have an interesting summary of the sort of "spin" found in abstracts that misrepresent study findings.  These include:

 • No acknowledgment of nonstatistically significant primary outcome
 • Claiming equivalence when results failed to demonstrate a statistically significant difference
 • Focus on positive secondary outcome
 • Nonstatistically significant outcome reported as if they were significant

...and several others.

"Misrepresentation of Randomized Controlled Trials in Press Releases and News Coverage: A Cohort Study"

Wednesday, September 12, 2012

When Positive D-Dimers are Negative

This is the latest article from Jeff Kline, published in Thrombosis and Haemostasis (don't you all subscribe to that, too?), concerning pulmonary embolism and d-Dimer.

Wouldn't it be great if the d-Dimer wasn't a dichotomous cut-off?  Where, if a patient were of sufficiently low pre-test probability, a d-Dimer value that was nearly negative still contributed adequately to a negative likelihood ratio to reduce the probability of a significant pulmonary embolism?  Well, that's the theory behind this article – which looks at d-Dimer measurements combined with age, Wells' score, and Revised Geneva scores.

There are a lot of complex tables in this article breaking down the various potential cut-off values for d-Dimer along with different pre-test probabilities, and the concept presented is that potentially higher cut-off values of d-Dimer can be used without missing PEs larger than sub-segmental.  This is presented in context that a higher cut-off might allow reductions in imaging, which seems fair.

However, the most interesting thing in this article to me is Figure 3 – which is d-Dimer concentration compared with fractional obstruction of pulmonary vascular tree.  It is, unfortunately, pretty clear there's not a great linear relationship between dimer and pulmonary obstruction.  Most low d-Dimers had < 5% obstruction of the vascular tree, but at least one patient with a "negative" d-Dimer had 20% obstruction.  Beyond that, patients were just as likely to have 90% obstruction with modestly elevated d-Dimers than with massively elevated d-Dimers.

"D-dimer threshold increase with pretest probability unlikely for pulmonary embolism to decrease unnecessary computerized tomographic pulmonary angiography"

Monday, September 10, 2012

Not-So Routine Surgery on Dabigatran

This correspondence, published in Blood in March, was probably pretty easy to overlook.

A patient enrolled in RE-LY, the trial comparing dabigatran and warfarin for non-valvular atrial fibrillation, underwent open aortic valve replacement surgery.  As instructed, he discontinued his dabigatran two days prior to the surgery.

Had a little bit of a bleeding problem.

After 26 units of RBCs, 5 packs of platelets, 22 units of FFP, 5 x 10 units of cryoprecipitate, two doses of protamine, two doses of tranexamic acid, and five doses of Factor VIIa, the patient was finally stable enough to be evacuated to the ICU for dialysis to remove the remaining dabigatran.

What's most fabulously ironic about this correspondence is that the authors use this horrifying case to sprightly conclude Factor VIIa and hemodialysis are viable and effective reversal strategies for dabigatran-associated bleeding.

The patient – "The postoperative course was complicated by prolonged ventilation/Enterobacter pneumonia, asymptomatic nonocclusive femoral DVT (by surveillance ultrasonography [postoperative day (POD 7)]), and acute-on-chronic renal failure. Discharge to a rehabilitation facility occurred on POD56." – probably disagrees.

Would you be surprised if I mentioned there's a COI issue involving the authors and the manufacturer?

"Recombinant factor VIIa (rFVIIa) and hemodialysis to manage massive dabigatran-associated postcardiac surgery bleeding"

Saturday, September 8, 2012

Dabigatran - It's Everywhere!

Dabigatran, you may know it at Pradaxa, the first in a string of potential blockbuster oral anticoagulants, has a few problems.  Lack of effective reversal options, poor prescriber understanding of drug-drug and GFR interactions, and reduced dosing options should make physicians wary of this medication.

Well, it's not.

This is a dataset gleaned by an ongoing physician audit covering ~4800 U.S. practioners between 2007 and 2011, used to estimate prescription trends for the U.S.  Warfarin prescriptions were reasonably stable between 2007 and 2010, but then have dropped approximately 20% over the course of 2011.  The medication taking up the slack?  Dabigatran.

If you accept the findings from RE-LY, then you're probably OK with its use in non-valvular atrial fibrillation.  Unfortunately, 37% of the prescriptions were off-label, outside the FDA approved indications. Then, within the remaining 63%, there's no breakdown for whether it was valvular or non-valvular atrial fibrillation.  So, the percentage of off-label use is probably even higher than found in the data.

It would seem to be prudent to be cautious with a new medication that's already being investigated by the FDA for serious bleeding complications.  Luckily for the manufacturer, that's not happening, and prescription expenditures for dabigatran already exceed those for warfarin – over $160 million per quarter.

"National Trends in Oral Anticoagulant Use in the United States, 2007 to 2011"

Thursday, September 6, 2012

More Failed Therapies for Sinusitis

For routine, office-based diagnoses of acute sinusitis, we've seen that antibiotics are unlikely to be beneficial.  The other theory behind treatment is attenuation of the inflammatory response, promoting sinus drainage.  Intranasal steroid sprays have inconclusive data.  This is a trial of systemic steroids, theorizing that intranasal steroids suffer from inadequate tissue penetration.

There are a lot of issues with this trial.  Whether it's clinically significant or not, the 30mg/day dose of prednisolone is below the typically used doses of 50mg or 60mg.  There were 54 treatment locations and 68 family physicians involved in this study over a 2 1/2 year period – and only managed to enroll 185 patients.  For a problem "frequently encountered" in primary care, it's a little hard to have confidence there aren't biases present with enrollment.

The authors followed many different clinical outcomes, as well as the SNOT-20 score, at several different time points, and the easiest way to sum it up is to say there are probably no clinically relevant differences between groups.  The trends nearly all favored prednisolone, but the absolute differences in outcomes provided NNT between 10 and 33.  A larger trial might have detected a statistically significant benefit to steroids – or it might not – but most enrolled patients had symptom improvement, regardless.

"Systemic corticosteroid monotherapy for clinically diagnosed acute rhinosinusitis: a randomized controlled trial"

Tuesday, September 4, 2012

It's Too Hot To Fight & Other Fables

There's a mythology regarding temperature and violent crime – both increase in tandem up until a certain point, at which it becomes "too warm".  This study, a retrospective analysis of violent crime from a six-year period in Dallas, TX, generally confirms the increase in violence as the temperature increases.

The authors additionally propose, however, a curvilinear relationship based on the data that interprets an inflection point at 80-89 degrees a bit aggressively, considering they only have one data point above 80-89 with which to define the further trend.  The absolute differences between total numbers of violent assaults in each temperature bracket are small enough, it's a little hard to confidently say there's a point at which it becomes too hot for violent crime.  It makes sense, of course, but that's more editorializing.

Perhaps they could attempt to externally validate these findings in Iraq – which seems awfully hot and violent.  They also note there is a strong correlation between temperature and hours of daylight – but it seems as though that'd be rather difficult to control for one or the other.

And, tying this entire issue into climate change is another unusual matter entirely....

"Temperature and Violent Crime in Dallas, Texas: Relationships and Implications of Climate Change"

Saturday, September 1, 2012

The End of IABP?

Adding to the "don't do anything, just stand there!" file, another relatively frequently used cardiovascular support tool – intra-aortic balloon counterpulsation – might be on the chopping block.

Typically used in cases of severe cardiogenic shock secondary to acute myocardial infarction, IABP is used to reduce strain on the stunned myocardium.  The first IABP-SHOCK pilot of 45 patients showed no mortality difference, but a significant improvement in BNP levels with IABP use.  This is the follow-up study, enrolling 600 patients to IABP or best available medical therapy.

Both groups were similarly ill – the IABP group had 6% more anterior STEMIs – and had nearly identical outcomes.  There were 1.5% more survivors in the IABP group, but the p value was 0.69.  Adverse events were similar – although the control group tended towards increased sepsis, which seems a little odd.  There was an expected random assortment of subgroups favoring one therapy or another, but nothing that would seem to be specifically hypothesis generating.

In the end, the authors rather grimly state that, despite some surrogate markers appearing to be improved in the IABP group, there is no evidence to support routine use of IABP in cardiogenic shock secondary to acute myocardial infarction.

"Intraaortic Balloon Support for Myocardial Infarction with Cardiogenic Shock"

Thursday, August 30, 2012

FFR(CT) - So Close, Yet So Far

As David Newman will tell you, the additive prognostic and predictive value for stress tests is fairly weak.  CT coronary angiograms are still a test looking for the correct population.  Conventional coronary angiography is expensive, invasive, and clearly not appropriate for the massive population of low risk patients we evaluate.

So, how about a non-invasive test that combines three-dimensional anatomic coronary reconstruction with predictive flow dynamics to identify lesions resulting in ischemia?  This test is CT coronary angiography combined with computed fractional flow reserve.  And, it would be a beautiful thing if it were ready for primetime – but it's not.

I've reviewed a previous trial of FFR(CT).  This is a larger study, published in JAMA, of 285 patients with suspected CAD who underwent CCTA with FFR(CT), followed by conventional coronary angiography with invasive FFR measurement.  Figure 1 summarizes the results relatively succinctly – but essentially, 56 of the 172 FFR(CT) patients with lesions calculated as ischemic were false positives.  On the flip side, 67 of the 80 FFR(CT) patients with lesions calculated as non-ischemic were false negatives.  The per-vessel performance of FFR(CT) basically added no additional diagnostic AUC to CT alone.

The study is sponsed by HeartFlow, and authored by several physicians disclosing conflicts of interest with diagnostic imaging manufacturers.  Unsurprisingly, the authors try to spin the positive out of it in their conclusions and abstract.

"Diagnostic Accuracy of Fractional Flow Reserve From Anatomic CT Angiography"

Wednesday, August 29, 2012

Pop & Suture Abscesses Closed

Upending another slice of traditional dogma, brought to my attention by Andy Neill, this is a systematic review and meta-analysis of the 7 randomized clinical trials comparing primary closure of cutaneous abscesses with secondary.  I love articles that challenge routine practice – some of which was actually transcribed from stone tablets into Tintanelli by Moses.

Unfortunately, as weak as the evidence may be for packing abscesses, antibiotics with abscess, etc., the evidence from this meta-analysis really is only serviceable as underpinnings to justify further trials revisiting standard practice.  The 915 patients included in this meta-analysis were primarily anogenital abscesses drained in an operating room environment by surgeons and many received antibiotics.  Some of the outcomes measured in this study make sense and are probably generalizable – healing time and time off work – which obviously will favor the patients with primary closure.  The less generalizable is the 600 patient subset which tracked abscess recurrence, which also has a ton of heterogeneity between studies.

Is it reasonable to perform some sort of abscess closure?  I think it probably is – depending on the amount of potential disfigurement, there's probably a discussion of risk/benefit that can be had.  There are also probably varying techniques of suturing that could be entertained, loosely approximating some part of the abscess, perhaps with a wick, rather than tightly re-approximating skin edges.

The authors state they are undertaking their own randomized trial.

"Primary closure of cutaneous abscesses: a systematic review"

Monday, August 27, 2012

Who's Burned Out? We Are!

In a survey of U.S. physicians, published in Archives of Internal Medicine, Emergency Physicians hold the dubious honor of being the most "burned out" specialty.  This estimation of burnout is based on survey questions regarding emotional exhaustion, depersonalization, and personal accomplishment.  Unsurprisingly, dermatology was not a terribly burned out specialty – likely because they also ranked quite highly in time for personal and family life.  Surgical specialties, despite ranking at the bottom for family life, were only in the middle of the road for burnout – probably indicating the nature of the work plays a role in the strain.  Self-selection bias always plays a role in these surveys, of course, considering the response rate was only 26%.

Compared with the employed general population, physicians were more burned out and had less time for family – which may or may not be related to the 25% more hours per week worked, although, there were many differences between the physician cohort surveyed and the comparison.  

The New York Times discusses this article and uses a vignette of a "missed diagnosis" as symptomatic of the disruption of quality of care due to burnout.  While it may be true that burnout relates to healthcare quality, the specific case presented seems to fall more into a category of reasonable conservative management of the most likely condition, with appropriate further enquiry made at a re-visit due to persistent symptoms.

"Burnout and Satisfaction With Work-Life Balance Among US Physicians Relative to the General US Population"

Friday, August 24, 2012

Head to Head With Head CT Rules

The "headline" you'll see from this article is that the Canadian Head CT Rule outperforms the New Orleans Criteria for radiographic imaging in minor head trauma.  Specifically, it outperforms it in this prospective, observational cohort from several hospitals in Tunisia, consecutive patients with blunt trauma to the head and at least one symptom secondary to the head trauma.

The most striking thing about this article, however, remains the gruesome number of false positives generated by each of these head CT decision rules.  While, obviously, the intent is to capture all the cases requiring neurosurgical intervention, the New Orleans Criteria could not rule out potential need for neurosurgical intervention in 1,180 out of 1,582.  When the theoretical purpose of these rules is to prevent "scanning everyone", we're not getting much bang for our buck.  The Canadian Head CT Rule was better – but still indicated a need for scan in 656 out of 1,582.

While the article focuses mostly on the need for neurosurgical intervention in GCS 15 patients, it's interesting to see their "secondary outcomes" which did not need "intervention".  Only 34 total patients in their cohort required intervention – while they found 133 skull fractures, 41 subdurals, 45 epidurals, 69 subarachnoids/hemorrhagic contusions, and 1 case of pneumocephalus.  The Canadian rule would have missed 11 of the 218 "clinically significant" findings, for a sensitivity of 95%.  The article does not specific precisely which types of findings were missed, but, clearly, many of those may be argued to be not significant.  Unfortunately, deriving a better rule based on a more liberal definition of "clinical significance" is likely to result in more missed interventions – but it's still probably worth trying.

"Prediction Value of the Canadian CT Head Rule and the New Orleans Criteria for Positive Head CT Scan and Acute Neurosurgical Procedures in Minor Head Trauma: A Multicenter External Validation Study"

Wednesday, August 22, 2012

How Fast Can We Rule-Out AMI?

Six hours?  Two hours?  One hour?  McDonalds' drive-thru?

This is the paper from Archives of Internal Medicine that's been making the rounds in the lay press regarding how quickly the ER should be able to detect your AMI with the new highly-sensitive troponins.  This is the APACE, prospective, international, multi-center study evaluating patients with "symptoms suggestive of acute myocardial infarction" and onset within the last 12 hours.

In this cohort, 1247 patients were recruited – and >300 were excluded for either going straight to the cath lab or having "another procedure performed" at the 1-hour time mark – and received hs-cTnT at index, 1, 2, 3, and 6 hours after presentation.  Myocardial necrosis was defined as a hs-cTnT >99th percentile, which for this assay is 14 ng/L, and a diagnosis of acute MI was made by two independent cardiologists upon review of records and lab results.

The authors split their cohort into two groups, a derivation cohort and a validation cohort, and did some statistical wrangling to come up with two cut-off strategies – one for rule-in and one for rule-out.  They were able to make diagnostic decisions on ~76% of their cohort at the one-hour time point, and 52 ng/L at presentation or an increase within an hour of 5 ng/L or more was ~94% specific for AMI.  Likewise, 12 ng/L and an increase less than 3 ng/L at 1 hour was ~100% sensitive for AMI.  The remaining 25% of their cohort was in a non-diagnostic zone.  At 30 days, there was one death in their rule-out cohort, for a 99.8% survival rate.

So, can you use this strategy?  If you feel as though this study is externally valid to your populations and you're using the same Roche Diagnostics test, you certainly may.  Every piece of data is something you can incorporate to your discussions with a patient regarding diagnostic certainty and risk.  Even an extra hour occupying an ED bed rather than moving out to a chest pain observation facility can significantly impede ED flow, while observation admissions are costly and inconvenient to patients.  The ideal strategy will depend on the capabilities of individual departments.

This study, along with the primary author, are sponsored in part by Abbott, Roche, and Siemens.

Monday, August 20, 2012

Intensive Blood Pressure Control in ICH

Not much works to treat ICH – and this retrospective analysis of INTERACT1 tries to coordinate a couple leaps to pull a different spin out of old data.  Unfortunately, it's still a re-analysis of essentially a negative trial, and that limits its utility for the purists.

INTERACT1 randomized patients with ICH to either "intensive" blood pressure lowering or conventional treatment – most of whom received some BP control – and noted significant improvements in hematoma volume.  However, the 90-day safety/clinical outcomes data did not show any difference in mortality or dependency.  This publication reverses the strategy, taking a look at the associations between good clinical outcomes and hematoma volume – and finds that increases in hematoma volume at 24 hours clearly predict poor clinical outcomes.

So, if intensive BP control reduces hematoma expansion and reduced hematoma expansion improves clinical outcomes, then why was INTERACT1 negative?  I suspect we'll find out more when INTERACT2 is published....

"Hematoma growth and outcomes in intracerebral hemorrhage: The INTERACT1 study"

Sunday, August 19, 2012


Medication shortages are affecting many hospitals – we're low/out of prochlorperazine, injectable metaclopromide, etomidate, propofol, brevital – and one of the replacements we've recently been introduced to is "Propoven", a European manufacture of propofol.

It has only minor differences from propofol, but it should be noted it requires strict sterile technique when handling and has more medium-chain fatty acids.  An informational letter from Kabi describing a few of the differences is here:

AHRQ Infection Control @ ACEP

Jeremiah Schuur, featured on EM Lit of Note for his timely critique of the inadequacy of the "quality" measure for non-contrast head CT, passes along a notification of a pre-ACEP conference in ED infection control practices.

Sponsored by AHRQ, ACEP, and infection control societies, find more information about the conference here: