More Tests, Longer Turnaround, Longer LOS

In this Friday’s edition of “It’s Science!”, we cover this recent publication demonstrating, essentially, what we already knew:  throughput suffers when test turnaround times are longer!

To do so, however, requires (apparently) cross-classified random-effect modeling, linking Emergency Department information systems to laboratory test data.  These authors evaluated a retrospective, multi-site cohort consisting of 27,656 linked ED and laboratory encounters and modeled the attributable effect of test turnaround time on ED length-of-stay.  They discovered, rather obviously, patients receiving more tests had longer ED LOS.  Working backwards, furthermore, they found for every 30 minute increase in laboratory test turnaround time, there was an approximately 17 minute increase in median ED length-of-stay.  It’s not a 1:1 relationship – as you can imagine situations where the ED LOS is dictated rather by radiography, procedures, or consultations, rather than laboratory testing – but increases in a relatively linear fashion.

So, depending on the structure and flow of your Emergency Department, there may be substantial benefits to focusing on improved laboratory turnaround times.  And, likewise, you can probably improve all your times by simply ordering fewer tests!

“The Effect of Laboratory Testing on Emergency Department Length of Stay: A Multihospital Longitudinal Study Applying a Cross-classified Random-effect Modeling Approach”

Ceftriaxone: It’s Not For Stroke

Negative trials are just as important – if not moreso – than positive trials.  Without negative trials, well-meaning clinicians continue with unproven and unverified treatments, incurring unnecessary costs and exposing patients to unneeded potential adverse effects.

This prospective trial of 2,550 patients randomized patients post-stroke to four days of prophylactic ceftriaxone versus no treatment, using an open-label, masked endpoint design.  The theory – some of the mortality and disability post-stroke may be in excess as result of infectious (primarily respiratory) complications.  The fact – no.  There was no difference in functional outcome at any modified Rankin Score ordinal cut-off, nor mortality.  There were actually not fewer cases of pneumonia in the ceftriaxone group, but urinary tract infections were significantly suppressed.  Clostridium difficile infection occurred only in two cases, both in the ceftriaxone group.

As a random aside, the median NIHSS in this trial was 5 – and 37% of patients achieved mRS 0-1.  This is rather surprisingly low, considering the SITS-MOST cohort and pooled placebo groups from RCTs with median NIHSS’ ~12 had achieved mRS 0-1 in 38% and 33% of the time, respectively.  A window into the lack of generalizability of the thrombolysis RCTs?  Perhaps – or just a curiosity.

Oh, but, back to the point – it does not appear to have clinical utility to routine use ceftriaxone as antibiotic prophylaxis after acute ischemic stroke.

“The Preventive Antibiotics in Stroke Study (PASS): a pragmatic randomised open-label masked endpoint clinical trial”

Should Men and Women Use Different Troponin Cut-Offs?

Much ado is made regarding potential differences in symptoms between men and women presenting with acute coronary symptoms.  Little is mentioned, however, about potential differences in laboratory thresholds between the sexes.  Considering women, on average, have decreased myocardial mass than men, any ischemic insult simply damages a smaller absolute quantity of myocardium.  Less damaged tissue, then, ought to lead to lower circulating biomarkers.

Why haven’t we tried this before?  Because the limit of detection of conventional troponin assays are above the clinically important thresholds for delineating such small quantities of circulating molecules.  However, with the advent of highly-sensitive troponins with reasonable precision below  the conventional troponin cut-off of 50 ng/L, it’s now a reasonable concept for investigation.

These authors conducted a yearlong prospective evaluation of all patients with suspected acute coronary syndrome, collecting conventional and highly-sensitive troponins on each.  Treating clinicians and initial adjudication of myocardial infarction were blinded to the results of the hsTnI.  Following conclusion of the study, records and unmasked hsTnI values were provided for independent adjudication and diagnosis changes accordingly.

Initially, 19% of men were diagnosed with Type 1 MI based on conventional troponin testing.  After using a gender-specific cut-off for men of 34 ng/L, only a handful of additional cases were re-classified – rising to 21%.  For women, 11% were initially diagnosed with Type 1 MI.  Using a gender-specific cut-off for hsTnI of 16 ng/L, however, doubled the diagnosis cohort to 22%.

Of course, simply lowering the threshold for any assay increases the rate of diagnosis.  In order to answer the question of whether the re-classified cases were clinically appropriate, all patients were also followed for survival free from death or MI.  While women not diagnosed with MI at initial presentation did well throughout the follow-up period, the women reclassified as MI using the hsTnI threshold suffered the same dismal outcomes as those initially diagnosed with MI.

I like this concept, and this is promising preliminary data.  It remains to be seen whether treatment, including increased treatment intensity for women, based on the gender-specific cut-offs changes clinical outcomes – or whether splitting these little nanograms worth of hairs is just overdiagnosis.  The good news: a clinical trial is ongoing.  I look forward to their results.

“High sensitivity cardiac troponin and the under-diagnosis of myocardial infarction in women: prospective cohort study” (free fulltext)

Unreliable Information About Drug Use? In the ED? Never!

This simple observational series illuminates the likely truths behind our anecdotal experience – patients are either clueless or deliberately misleading regarding their ingestion of foreign substances.

For the purposes of this investigation, “drug use” is not restricted to illicit substances – these authors also explored the reliability of reporting of prescription medications.  In this prospective, year-long enrollment, 55 patients were selected randomly from a larger cohort to have a urine sample submitted for liquid chromatography/mass spectrometry.  The LC/MS assay utilized was capable of detecting 142 prescriptions, over-the-counter drugs, and drugs of abuse and their metabolites.  A drug whose level of detection was lower than 2 half-lives was reported as discordant with patient self-reporting.

All told, 17 out of 55 patients provided accurate medication histories, based on those detected on LC/MS.  Over half the patients under-reported – including a patient with 7 unreported drugs detected – while 29% over-reported a drug not subsequently detected.  Interestingly, illicit drugs were the least likely to be mis-reported, although, that may simply be a reflection of the higher prevalence of prescription and OTC medications.

Such observations are limited by the accuracy of the assay utilized, which has not been validated.  However, it ought come as no surprise many patients either intentionally or unintentionally misrepresent all possible drug exposures.  While not all omissions are clinically relevant, certainly, non-compliance and misinformation may have important implications for diagnosis and treatment.

“The Accuracy of Self-Reported Drug Ingestion Histories in Emergency Department Patients”

Prednisone … for Pneumonia?

The utility of antibiotics for eradication of bacterial pathogens from the lower respiratory tract is a given.  Use of steroids – also known for their immunosuppressive properties – not so much.

But, one can imagine clinical utility for steroids in acute infection.  Not every function of the immune system results in desirable patient-oriented effects.  Immunologic host responses include release of many inflammatory cytokines responsible for organ dysfunction, and steroids are already part of accepted therapy for several specific manifestations of pneumonia.  Based on prior results in smaller trials, these authors suspected use of steroids might be of benefit – both in mortality and in time to symptom resolution.

With 785 patients allocated in blinded fashion to 50 mg of prednisone daily or placebo, patients receiving prednisone reached “clinical stability” in a median of 3 days, compared to 4.4 days for the placebo cohort.  Hospital length-of-stay was reduced to 6 days from 7, and intravenous antibiotic use was cut by a day.  There were few important adverse effects overall, and the only consistent harm apparent in these data was increased hyperglycemia associated with corticosteroid use.

The accompanying editorial in The Lancet states adjunctive therapy with steroids is a therapy whose time has come, based on healthcare savings due to resource utilization.  In the context of other published studies, this observed reduction in time to vital sign normalization is valid.  However, whether the effect of steroids is truly beneficial or akin to simply masking the underlying clinical state by suppression of pro-inflammatory cytokine release is less certain.  Use of anti-pyretics blunts outward signs of systemic inflammatory response syndrome, and beta-blockers likewise reduce the tachycardia resulting from physiologic stress without specifically treating the underlying process.  It is hard to associate the outcomes measured in this trial with actual expedited clinical cure.

Reductions in length-of-stay and IV antibiotic use are reasonable patient-oriented and system-oriented outcomes, however, so the decision ultimately rests with the magnitude of harms – and the harms are certainly real.  Previous studies have suggested increased early recurrence or persistence of pneumonia, in addition to uncontrolled hyperglycemia.  These authors hoped to measure a 25% reduction in mortality – which was a bit of an odd expectation, given the ~2-4% expected absolute mortality – and no such suggestion of benefit was observed.

Simply put, this is not ready for prime-time or guideline-level adoption.  It is certainly worthy of further study, but steroids should not be used routinely outside the scope of prospective monitoring.

“Adjunct prednisone therapy for patients with community- acquired pneumonia: a multicentre, double-blind, randomised, placebo-controlled trial”

NIHSS Scores are Not Created Equal

This is hardly news to anyone with a clinical practice, but it’s a topic rarely addressed in stroke trials – that patients with identical NIHSS can have a wide range of downstream disability.

This is a retrospective analysis of the VISTA registry, which collates non-thrombolysis acute stroke trial data, and is generally useful for identifying predictors of long-term prognosis and outcomes.  These authors used six hypothesized “profiles” of stroke syndromes with distinct constellations of disabilities, and matched a total of 10,271 patients from their database to one of the six.  Using their most disabling stroke subtype profile as reference, the authors noted three different syndromes – with median NIHSS 10, 9, and 7 – all had similar likelihood of favorable outcomes.  However, even though the NIHSS and good outcomes were similar, the disabilities and clinical profile associated with one of these cohorts translated to twice as likely to be deceased at 90 days.  In essence – similar “numbers”, but very different outcomes.

There’s nothing here usable for direct knowledge translation – but, it does hearken back to my oft-repeated statements regarding the heterogeneity of stroke syndromes, outcomes, and likelihood of benefit or harm from pharmacologic revascularization.  Quite simply, data sources such as this – and those including patients from thrombolysis trials – ought be better utilized to predict patient-specific outcomes.

“National Institutes of Health Stroke Scale Item Profiles as Predictor of Patient Outcome”

You’re Damn Screwed on Ticagrelor

If you’re unlucky enough to suffer intracranial bleeding, your unluckiness is compounded if you’re concurrently taking any sort of anticoagulation.  Some agents have relatively-effective reversal options – typically prothrombin concentrate complexes or fresh frozen plasma.  Anti-platelet agents, however, tend to irreversibly bind and inactivate platelets – and the only theoretical reversal strategy is transfusion with new, unblemished platelets.

That definitely won’t work for ticagrelor.

This brief letter in the NEJM details the unfortunate case of a man suffering a hemorrhagic transformation following a stroke while on ticagrelor.  As part his treatment, these authors transfused the patient a total of 17 units of platelets.  After said transfusion, the platelet-reactivity index barely rose from 0% to approximately 10% immediately following the transfusion – but then returned to 0% an hour later, and remained unchanged at 0% seven hours later when rechecked.  As you might expect, with 0% platelet activity, the patient expired as a result of his intracranial bleeding.

If you can manage not to waste blood products in such futile action, please do so.  Also, beware!

“Inefficacy of Platelet Transfusion to Reverse Ticagrelor”

The Wholesale Revision of ACEP’s tPA Clinical Policy

ACEP has published a draft version of their new Clinical Policy statement regarding the use of IV tPA in acute ischemic stroke.  As before, the policy statement aims to answer the questions:

(1) Is IV tPA safe and effective for acute ischemic stroke patients if given within 3 hours of symptom onset?
(2) Is IV tPA safe and effective for acute ischemic stroke patients treated between 3 to 4.5 hours after symptom onset?

Most readers of this blog are familiar with the mild uproar the previous version caused, and this revision opens by stating “changes to the ACEP clinical policies development process have been implemented, the grading forms used to rate published research have continued to evolve, and newer research articles have been published.”  Left unsaid, in presumably a bit of diplomacy, were the conflicts of interest befouling the prior work.  Notably absent from this work is any involvement from the American Academy of Neurology.

What’s new, with a new methodology-focused rather than conflicted-expert-opinion approach?  Most obviously, there’s a new Level A recommendation – focused on the only consistent finding across all tPA trials: clinicians must consider a 7% incidence of symptomatic intracranial hemorrhage, compared with 1% in the placebo cohorts.  The previously Level A recommendation to treat within 3 hours has been downgraded to Level B.  Treatment up to 4.5 hours remains Level B.  Finally, a new Level C recommendation includes a consensus statement recommending shared decision-making between the patient and a member of the healthcare team regarding the potential benefits and harms.

Most of the reaction on Twitter has been, essentially, a declaration of victory.  And, in a sense, it is certainly a powerful statement regarding the ability for like-minded patient advocates and evidence purists to coalesce through alternative media and initiate a major change in policy.  To critique this new effort is a bit of punishing the good for lack of manifesting perfect, but there are a number of oddities worth providing feedback to the writing committee:

  • The authors provide a curious statement:  “The 2012 IV tPA clinical policy recommendation to ‘offer’ tPA to patients presenting with acute ischemic stroke within 3 hours of symptom onset was consistent with other national guidelines. Unfortunately, the essence of the term ‘offer’ may have been lost to readers and has therefore been avoided in this revision.”  I rather find “offer” a lovely term, in the sense it expresses a cooperative process for proceeding forward with a mutually agreed upon treatment strategy.  Rather than discard the term, clarification might have been reasonable.
  • They mention ATLANTIS as Class III evidence with regard to the 3-4.5 hour question.  I can see how its classification may be downgraded given the multiple protocol revisions.  That said, its inability to find a treatment benefit in spite of extensive sponsor involvement ought be a more powerful negative weighting than currently acknowledged.  Given the biases favoring the treatment group in ECASS III (given a Class II evidence label), the cumulative evidence probably does not support a Level B recommendation for the 3-4.5 hour window.
  • One of my Australian colleagues in private communication brings up a small letter from Bradley Shy, previously covered on this blog, mentioning a statistical change to ECASS III.  This statement could acknowledge this post-publication correction and its implications regarding the aforementioned imbalance between groups.
  • The authors fail to acknowledge the heterogeneity of acute ischemic stroke syndromes and patient substrates, and the utter paucity of individualized risk or benefit assessment tools – in no small consequence of the small sample sizes of the few trials rated as Class I or Class II evidence.  This is a powerful platform with which to state clinical equipoise exists for continued placebo-controlled randomization.  As we see from the endovascular trials, the acute recanalization rate of IV tPA is as low as 40% – with many patients re-occluding following completion of the infusion.  Patients need to be selected less broadly with respect to likelihood of benefit compared with supportive care.  I believe tPA helps some patients, but it should be a goal to dramatically reduce the costs and collateral damage associated with rushing to treat mimics and patients without a favorable balance of risks and benefits.  For these authors to recommend treatment in “carefully selected patients” and “shared decision-making”, more guidance should be provided – and absent the evidence to support such guidance, they should be calling for more trials!

The comment period is open until March 13, 2015.

“Clinical Policy: Use of Intravenous tPA for the Management of Acute Ischemic Stroke in the Emergency Department DRAFT”

Addendum 01/18/2015:
The SAEM EBM interest group is compiling comments on the evidence for feedback to the SAEM board of directors.  These are my additional comments after having had additional time to digest:

  • I agree with sICH as a Level A recommendation.  Both RCTs and observational registries tend to support such a recommendation.  Whether the pooled risk estimates are usable in knowledge translation to individual patients is less clear.  The risk of sICH is highly variable depending on individual patient substrate.  There are several risk stratification instruments described in the literature, but none are specifically recommended/endorsed/prospectively validated in large populations.
  • It is uncertain regarding the NINDS data whether their intention is to present pooled Part 1 and Part 2.  The prior clinical policy used only Part 2 for their NNT calculation, giving rise to an NNT of 8 instead of 6.  It appears they are pooling the data from both parts here.  Either is fine as long as it’s explicitly stated – the primary outcome differed, but the enrollment and eligibility should have been the same.
  • ECASS seems to be missing from their evidentiary table.  The ECASS 3-hour cohort data is available as a secondary analysis.  However, such would probably be Class III data of no real consequence for the recommendation.
  • Level B is probably an acceptable level of recommendation for tPA within the 0-3 hour window.  “Moderate clinical certainty” is reasonable, mostly on the strength of the Class III data.  However, the “systems in place to safely administer the medication” is not clearly addressed in the text.  Most of the published clinical trial and observational evidence involves acute evaluation by stroke neurology.  Does the primary stroke center certification practically replicate the conditions in which patients were enrolled in these trials/registries?  Perhaps this should be split out into a separate recommendation regarding the required setting for safe/timely/accurate administration.
  • Level B is difficult to justify for the 3 to 4.5 hour time window.  There is Class II evidence from ECASS III (downgraded due to potential for bias) demonstrating a small benefit.  The authors then cite Class III trial evidence from IST-3 and ATLANTIS in which no benefit was demonstrated.  Then, they cite the individual patient meta-analysis having similar effect size to ECASS III – because many of the patients in that subgroup come from ECASS III.  Basically, there’s only a single piece of Class II evidence and then inconsistent Class III evidence, which doesn’t meet criteria state for a Level B recommendation (1 or more Class of Evidence II studies or strong consensus of Class of Evidence III studies).  
  • With both Level B recommendations, the authors also reference “carefully selected” patients, but do not cite evidentiary basis regarding how to select said patients other than listing the enrollment criteria of trials.  If the “careful selection” is strict NINDS or ECASS III criteria, this should be explicitly stated in the recommendation.
  • The Level C recommendations to have shared decision-making with patients and surrogates ought to be obvious standard medical practice, but I suppose it bears repeating given the publications regarding implied consent for tPA.  They mention two publications regarding review and development of such tools, but there is no evidence supporting their efficacy or effectiveness in use.  Frankly, calling them a starting point in such a heterogenous population is along the lines of the broken clock that’s right twice a day.  I would rather say their dependence on group-level data minimizes their practical utility, and clinician expertise will be the best tool for individual patient risk assessment.

Feel free to add your comment and I will incorporate them into my feedback to SAEM.

(Failing to) Identify Severe Sepsis at Triage

This is the holy grail of predictive health informatics in Emergency Medicine – instant identification of serious morbidity, with the theoretical expectation of outcomes improvement due to early intervention.

And, more than almost any condition, accurate early identification of severe sepsis remains elusive.

This is an observational evaluation of the “Australian Triage Scale” in combination with infectious keywords as a tool to identify and manage patients with severe sepsis.  Patients were enrolled at presentation to the Emergency Department, and ultimately followed from triage through their ICU stay – where a clinical diagnosis of severe sepsis was used as the gold standard for outcomes. However, of the 995 patients triaged through the Emergency Department and ultimately diagnosed with severe sepsis, only 534 were identified at triage.  The authors present various diagnostic characteristics for each level of the ATS with regards to acuity, and the AUCs for sensitivity and specificity range from 0.457 to .567 (where 0.5 is basically a coin-flip).  So, the authors’ presented rule-based mechanism is nearly as likely to be incorrect as correct.  I’m not exactly certain how they came to the conclusion “the ATS and its categories is a sensitive and moderately accurate and valid tool”, but I tend to disagree.

These data are consistent with our a priori expectation for these sorts of tools.  The patients who trigger such rules are generally so obviously severe sepsis such rule-based notifications occur after clinician identification, and are simply redundant and alarm fatigue.  Conversely, patients with severe sepsis going undiagnosed upon initial presentation do so because of their atypical nature – and thus tend to fall outside rigid, rule-based constructs.  E.g., computers are not physicians … yet.

“Identification of the severe sepsis patient at triage: a prospective analysis of the Australasian Triage Scale”

Overstated Benefit of “Compliance” with Massive Transfusion

A couple days ago, @karimbrohi drew a bit of attention to this article on Twitter:

Compliance with Massive Transfusion Protocol improves outcome: [Protocol followed- 10% mortality. Not followed – 60%]
— Karim Brohi (@karimbrohi) January 3, 2015

Massive transfusion protocols are, essentially, the standard of care in advanced trauma care.  Coordinated systems to produce timely quantities of appropriate blood products are nothing new.  However, the contemporary usage of MTP has been to describe a protocol with fixed ratio of product – usually approximating a 1:1 ratio of PRBCs and FFP, and some programs include platelets.

This small study retrospectively evaluated the survival of 72 consecutive MTP activations at their trauma facility.  Compliance with 13 quality measures associated with resuscitation and transfusion was 66% overall.  Mortality rates in the cohorts with <60% compliance, 60-80% compliance, and >80% compliance with quality measures were 62%, 50%, and 10%, respectively.  Thus – compliance saves lives!


Tables 4 and 5 compare the baseline characteristics between survivors and non-survivors – and, frankly, it’s hard to decisively say “compliance” made the difference.  Worse initial GCS, median ISS, and AIS head/spine were all significantly associated with poorer outcomes.  Then, as far as differences in “compliance”, there was actually little difference between survivors and non-survivors regarding actual issuing and receipt of blood products.  Rather, the differences were in the quality measures associated with specific lab work – and hypothermia correction, which suffered greatly in non-survivors, almost certainly because they were in the OR for heroic measures rather than in the ICU.  So, it’s rather difficult to reliably state the quality of care was lower for those who did not survive.

And certainly not to account for the entire magnitude of this 60% to 10% mortality advantage!

“Compliance with a massive transfusion protocol (MTP) impacts patient outcome”