Month: October 2014

Using Patient-Similarity to Predict Pulmonary Embolism

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

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

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

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

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

Does Cardiac Catheterization Help After OHCA?

Yes!

Probably.

It sure seems like it.

But, we still don’t really know for whom.

We’ve reviewed several of the prospective and retrospective studies regarding post-arrest cardiac catheterization on this blog over the years.  The general conclusion – the authors are very enthusiastic about their outcomes, but their comparison groups are invalidated by selection bias.  So, unsurprisingly, when a systematic review and meta-analysis of these studies is performed – the same critiques hold.

This review identifies 50 studies with sufficient reporting and design for analysis.  27 of these studies describe use in STEMI complicated by OHCA – and outcomes are largely excellent, compared to typical OHCA survival.  Good neurologic survival, in the 18 studies reporting such, averaged 68.4%.  There’s not much debate regarding STEMI complicated by OHCA – cardiac catheterization, when available.

The problem, however, arises when evaluating patients with OHCA and no clear cause for arrest.  There were 15 studies comparing outcomes with and without cardiac catheterization – and, overall, good neurologic outcome was present in 58% versus 35.8%, with and without cardiac catheterization, respectively.  However, 11 of these 15 studies were retrospective, and patients undergoing conservative management tended to have poorer prognosis at baseline and those who underwent cardiac catheterization tended to have more prominent ischemic changes on post-arrest ECG.

So, it’s another garbage-in, garbage-out sort of meta-analysis and review.  It cannot be used to support universal expansion of the target population for cardiac catheterization after OHCA, and tells us, essentially, what we already knew.  Clearly, some patients – particularly those for whom a culprit lesion is identified – benefit.  For the remainder, the treatment population remains unclear, particularly in the face of the extraordinary resource utilization.

“Cardiac catheterization is associated with superior outcomes for survivors of out of hospital cardiac arrest: Review and meta-analysis”

ARISE, and Cast Off the Shackles of EGDT

The sound you hear is a sigh of relief from Emergency Physicians and intensivists regarding the outcomes of the Australasian Resuscitation in Sepsis Evaluation (ARISE).

As ProCESS suggested, and as many have suspected all along, it seemed the critical intervention from Early Goal-Directed Therapy was the early part – and less the SCO2 monitoring and active management of physiologic parameters using dobutamine and blood transfusion.  Now, we have a second study, in addition to ProCESS, supporting the same general conclusions.

ARISE enrolled patients with confirmed or suspected sepsis, and either hypotension refractory to 1L crystalloid fluid challenge or a lactate level of 4.0 mmol/L or more.  31 centers randomized 1,600 patients to undergo either EGDT or “usual care”, which entailed routine local clinical practice, excepting measurement of SCVO2 was forbidden.  EGDT, however, was provided by specially coordinated teams to ensure all patients received the intervention.  The primary outcome was death from any cause within 90 days, powered to detect an absolute risk-reduction of 7.6%.

Baseline characteristics between the two groups were quite similar, few patients dropped out of each arm, and, finally, there was no difference in the primary outcome – 18.6% vs. 18.8% (does it matter which is which?)  Indeed, of all the outcomes measured, only two differed in statistically significant fashion: the EGDT cohort departed the Emergency Department 30 minutes more quickly, and the EGDT cohort received greater vasopressor support – attributable entirely to the use of dobutamine in 15.4% of patients vs. 2.6% in the usual care arm.

As expected, resource utilization unique to EGDT, of course, was different – more and different types of central venous catheters, more arterial catheters, and more frequent use of blood products.  And, as we’re seeing – all of this is unnecessary.  As with ProCESS, “usual care” has become EGDT, excepting these elements.  Both groups received substantial, early crystalloid resuscitation, early appropriate antibiotic coverage, and departed the Emergency Department to a critical care setting quite quickly.

EGDT receives credit for making us aware the impact early identification and intervention can have on mortality.  However, it is time to leave EGDT behind and identify new resuscitation targets and sensible strategies for achieving them.

“Goal-Directed Resuscitation for Patients with Early Septic Shock”
http://www.nejm.org/doi/full/10.1056/NEJMoa1404380

Blood is Good for the Injured Brain – Or is it?

A guest post by Dr. Andrew Kirkpatrick (@AskEMdoc), an Emergency Medicine resident at the University of Texas Medical School at Houston.

Hypoxia, of course, is lethal to vulnerable cells.  Erythropoeitin, in many small trials, has been shown to be neuroprotective after injury.  So, given these apparently obvious beneficial and synergistic treatments, the authors of this study set out to answer the question: What happens when a patient with Traumatic Brain Injury (TBI) is given blood and erythropoietin?
This is a randomized control trial with a factorial (2×2) design that tested Erythropoietin (Epo) versus Placebo and Transfusion Threshold of 7.0g/dL versus 10g/dL to determine if either of the above interventions conferred the benefit of improved neurologic recovery in TBI.  A total of 200 individuals with TBI were randomized into one of four groups using a block randomization strategy, sorting individuals into groups with and without transfusion thresholds or Epo administration.  Both the treatment team and the follow up personnel were blinded to Epo administration, but only follow up personnel could be blinded to transfusion threshold group.  Then, to make this study even further convoluted, the Epo dose and frequency was changed mid-stream due to safety concerns, dividing the Epo arm into the Epo1 and Epo 2 groups.   They also changed the study from a superiority trial to futility, and unusually selected 0.15 as their one-tailed alpha – a choice severely restricting their ability to reject the null hypothesis.
Determination of primary outcome data was completed utilizing the Glasgow Outcome Scale (GOS) using a variety of strategies, including in person and phone follow up.   For all comparisons – both transfusion thresholds and Epo – no statistically significant difference was detected.  Given their small sample, they may simply have been unable to produce a difference – as Epo seemed to potentially have some beneficial effect, though transfusion certainly showed no such signal.  Risk of death, infection, ARDS, and Cardiovascular complications including VTE were evaluated for as well, and both Epo groups and the 10g/dL transfusion group had significant increased risk of adverse events.  And, regarding resource utilization, the transfusion threshold group obviously consumed more blood products.
The study had several limitations including change in Epo protocol with 1/3 of the patients already enrolled, inability to blind clinicians to transfusion threshold, the aforementioned statistical limitations, and generalizability limitations owing to its enrollment at only two trauma centers.  Overall the results were unable to demonstrate benefit to either strategy – but were able to demonstrate definite harms.  Until further evidence is presented, it is prudent to continue conserving blood products and abstaining from giving ineffective, expensive medication.
“Effect of Erythropoietin and Transfusion Threshold on Neurological Recovery After Traumatic Brain Injury A Randomized Clinical Trial”

http://www.ncbi.nlm.nih.gov/pubmed/25058216

Preposterous Inpatient Antibiotic Overuse

It is one matter entirely to give antibiotics for self-limited bacterial or viral conditions.  It is another matter to regularly, simultaneously prescribe multiple, redundant antibiotics with overlapping coverage, excepting a few particular situations.

And, we are clearly using overlapping coverage far more than just a few particular exceptions.

This review of proprietary data from inpatient antibiotic use in 505 U.S. hospitals between 2008 and 2011 looked at three types of antibiotic overlap – that for MRSA, for anaerobic bacteria, and for ß-lactam therapy.  These authors noted 32,507 cases in which patients received at least two consecutive days of redundant, simultaneous antibiotics.  The largest offender, by far:  82,018 days in which patients received both intravenous metronidazole and intravenous piperacillin-tazobactam.  The majority of the remainder were also anerobic overlap, and the authors also cited over a thousand cases each of dual-MRSA therapy or dual-ß-lactam therapy.

Now, there are certain tissues in which vancomycin has poor penetration, and vancomycin-intermediate strains are increasing – so it’s unreasonable to say all dual-MRSA therapy is inappropriate.  The same applies to the dual-ß-lactam therapy, as double-coverage for pseudomonas and other MDR pathogens frequently requires at least initial redundant therapy.  However, I think this data still reasonably reflects an abundance of opportunity to curtail inappropriate antibiotics use.

The authors, mostly employed by a health services consulting company, also try to do a cost-analysis to quantify the scope of the redundant use.  Unfortunately, in each case, they assume the more expensive antibiotic is the redundant one – which inflates and exaggerates their estimates.  Presumably, this comes out of the need to subsequently promote their company’s services, and these numbers are best ignored.

But, we can, at least, do much better than our present state of affairs.

“Economic Impact of Redundant Antimicrobial Therapy in US Hospitals”
http://www.ncbi.nlm.nih.gov/pubmed/25203175