Wednesday is for Stroke

It had been a few weeks since I perused the recently published articles in Stroke – and there were so many: 1) not quite enough for their own angry post, but 2) worth noting, so, here we are:

Emergency Department Door-to-Puncture Time Since 2014: Observations From the BEST-MSU Study
https://www.ahajournals.org/doi/10.1161/STROKEAHA.119.025106

This one comes with the nonsensical medical lay press article entitled “Mobile stroke units get patients to hospital faster than ambulances“. No, mobile stroke units do not warp the fabric of space-time. They are bound by the same laws of physics and traffic as the rest of us.

What this study actually shows is the difference between door-to-groin-puncture time in patients arriving via the MSU versus regular EMS. The result: it saved about 10 to 15 minutes to have the pre-hospital neurologist evaluation in-person or via telemedicine. Considering the observational evidence regarding the fragility of collateral circulation within the first few hours of large-vessel stroke, this is obviously favorable – but the actual clinical effect of a few minutes can only be minimal at best.

Functional Outcome Following Stroke Thrombectomy in Clinical Practice
https://www.ahajournals.org/doi/10.1161/STROKEAHA.119.026005

A huge German registry of stroke thrombectomy provides some insights into the “real world” outcomes. These authors spend a little time comparing their outcomes to those observed in the meta-analysis of the trials, but this comparison is obviously challenging due to the entry and perfusion imaging selection criteria of the trials. The big takeaway, however: real world mortality was 28.6% compared with the trial mortality of 15.3%. These findings should prompt further investigation into strategies to reduce risk for death.

Blood Pressure and Outcome After Mechanical Thrombectomy With Successful Revascularization
https://www.ncbi.nlm.nih.gov/pubmed/31318633

This is just an observational series – with a lot of missing data – looking for any association between blood pressure and outcomes following mechanical thrombectomy. The general trend: higher is worse. Something like the normotensive range is associated with the best outcomes and fewest complications. However, such observations cannot be assumed to suggest improving blood pressure control will reduce the frequency of downstream complications. More likely, the blood pressure is rising as a result of the complications – intracranial hemorrhage and cerebral edema – and it is only hypothesis-generating at this stage to say strict control will improve outcomes.

Magnitude of Benefit of Combined Endovascular Thrombectomy and Intravenous Fibrinolysis in Large Vessel Occlusion Ischemic Stroke
https://www.ncbi.nlm.nih.gov/pubmed/31311465

Very few analyses truly spur a full rolling-of-the-eyes, but this is one of them: taking matched cohorts from NINDS and SWIFT-PRIME to conjure up some sort of quantification of the benefit of endovascular therapy in large-vessel occlusion. It isn’t so much the basic principle of the analysis with which I quibble – but the fact they are only able to roll with 80 patients each from their NINDS tPA, NINDS placebo, and SWIFT-PRIME cohorts. Of course, if you recall, large-vessel occlusions were not specifically identified in NINDS, so these authors are imputing their presence based on NIHSS and deficit patterns, which is hardly a reliable means of identification. Then, with only 80 patients in each cohort, the imprecision of each comparison is so great it’s virtually pointless to rely on these findings in the patient-facing shared-decision-making information graphics they created. Without reading too much between the lines regarding why this nonsense was ultimately published, it should be noted the lead author is Jeff Saver’s son.

Pulmonary Embolism in Syncope, Forever

For about the zillionth time – the Prandoni observations of 16% prevalence of pulmonary embolism in syncope are not not generalizable.

This is the BASEL IX study (BAsel Syncope EvaLuation Study) whose primary purpose is to gather observational data regarding the outcomes of patients with syncope. These authors pulled a chunk of their data to address the question of whether PE was highly prevalent among those with syncope, as well as to describe the timing of such any downstream “missed” diagnoses.

The data presented comes from 1,895 patients enrolled in the main study, 1,397 of whom were eligible via being clinically stable enough to provide informed consent, and in whom a D-dimer result was available. Patients underwent clinical assessment by treating physicians without influence by the study.

Overall, only 14% underwent imaging for PE, with PE ultimately found in 1.4% of the cohort at initial presentation. About half the remaining patients who did not undergo imaging had normal D-dimer results and were retrospectively calculated to have low pre-test probability. By logical conclusion, however, that means half the remaining cohort was either non-low pretest or had elevated D-dimer – but still did not undergo imaging.

The follow-up period was up to two years in 83% of this cohort, and only 12 events occurred in follow-up, all in the non-low/elevated D-dimer group not undergoing initial imaging evaluation. Eight of these were PEs and 4 were cardiovascular deaths. The good news: the earliest subsequent PE diagnosis was 55 days, and earliest death was almost a year later.

So, no, again, no – no need to pursue PE in syncope unless otherwise clinically indicated, even if the patient may be at incidentally elevated risk.

“Prevalence of Pulmonary Embolism in Patients With Syncope”
http://www.onlinejacc.org/content/74/6/744.full

If You Guessed “Definitely Not ACS” …

This little observational study, part of a larger troponin-centric evaluation, looked at the predictive value of clinician gestalt for acute coronary syndrome. This has been evaluated before – including by this same group – but this cohort is three times the size of their prior effort.

Of the 1,391 patient encounters included, 207 had an acute MI, and another 33 died or underwent coronary revascularization within 30 days. Only 60 patients actually fell into the category of “definitely not ACS”, and 3 of those turned out to actually have an AMI. However, adding an ECG and a troponin to the initial gestalt was ultimately 100% sensitive for acute MI said “definitely not” cohort.

The other end of the spectrum – the “definitely ACS” side – was similar, with gestalt requiring supplementation by troponin and ECG testing to confirm.

One of the authors’ takeaways: a label of “definitely not” isn’t safe enough to forgo troponin testing. However, this comes with a big caveat: the enrolled patient cohort was specifically chosen for the main study because the treating clinician judged they required evaluation for ACS. Thus, effectively by definition, “definitely not” is incompatible with the study population.

You should not use this study to justify evaluation with additional or definitive testing in those who are truly “definitely not” ACS – the cohort here was enriched by 60 year old patients with hypertension, hyperlipidemia, prior MIs, smokers, and diabetes, and it would truly be the exception for one of these patients to “definitely not” have ACS. The 28 year-old for whom you think “definitely not” can still be evaluated as you feel appropriate.

“Can emergency physician gestalt “rule in” or “rule out” acute coronary syndrome: validation in a multi-center prospective diagnostic cohort study”
https://www.ncbi.nlm.nih.gov/pubmed/31338902

CRP for COPD

If you follow this blog, you’ve probably read various critiques of the use of procalcitonin to guide antibiotic prescribing. Procalcitonin, a non-specific inflammatory marker, provides a small amount of informational value regarding the underlying etiology of infection, but my underlying criticism of its envisioned use is:

  • The baseline rate of antibiotic prescribing is so poor, and the likelihood of poor outcomes so low, a safe reduction in prescribing is guaranteed.
  • It provides about the same area-under-the-curve for predicting bacterial etiologies as C-reactive protein.
  • The pro-procalcitonin studies and contributions are effectively covered in the fingerprints of the manufacturers of the assay.

So, then, replace the above complaints with – well, mostly just the top one, because here we are with CRP doing the same things for which procalcitonin is advertised, and the apparent conflict-of-interest is turned down a few notches.

In this study, 86 primary care clinics in England and Wales randomized patients with a diagnosis of COPD and a clinical diagnosis of an acute exacerbation to use of point-of-care CRP testing versus usual care. Similar to those studies seen with procalcitonin, prescribers were provided guidance with respect to various CRP levels and recommendations for either prescribing, possible prescribing, or do not prescribe. The primary outcome and secondary outcomes were associated with receipt of any antibiotics, quality of life, and adverse health outcomes.

Over the course of two years, 649 patients were randomized to the two arms, with a handful of each failing to properly undergo initial study procedures. The prescribing rate at the index visit in the “usual care” group: 69.7%. The prescribing rate with CRP: 47.7%. A winner is CRP!

Except that 76% of patients had CRP less than the threshold at which antibiotics were recommended. Another 12% were in the “antibiotics maybe” group. Thus, nearly 90% of the entire cohort were suspected of having no or limited benefit to antibiotics – so, of course any safety margin to deprescribing would be satisfied. And, considering the baseline rate of prescribing was 70%, again, there is basically no possible way a stewardship intervention could fail.

The editorial accompanying this article is darkly amusing, stating “the findings from this study are compelling enough to support CRP testing as an adjunctive measure to guide antibiotic use in patients with acute exacerbations of COPD”. However, it also goes on to note these data hardly identify “which patients (if any) truly benefit from antibiotic therapy”(emphasis mine). Some trials testing 100% antibiotic prescribing vs. zero prescribing (e.g., placebo) have found minimal, or no, benefit. As with procalcitonin, our problem is a pervasive culture of over-prescribing, and ultimate answer is the same for CRP: we don’t need to introduce a marginally informative test into this low-stakes patient population, we simply need to snap out of our collective insanity.

“C-Reactive Protein Testing to Guide Antibiotic Prescribing for COPD Exacerbations”
https://www.nejm.org/doi/10.1056/NEJMoa1803185

Déjà vu: The New Cutting Edge Treatment of Migraines

It’s rimegepant again!

As featured in such posts as … yesterday’s … it’s the same calcitonin gene-related peptide receptor antagonist, but in an entirely separate large, multi-center, double-blind, randomized controlled trial.

And, the results are essentially the same. In this trial, 1,466 participants were randomly assigned to rimegepant or placebo for treatment of a single migraine headache of moderate-to-severe intensity. At two hours post-dose, response rates were 21% with rimegepant and 11% with placebo. Adverse events were similarly rare and generally mild.

Yet again, this tells us rimegepant is better than nothing – and in no way informs us regarding efficacy in their hypothetical patient population: those without response to either over-the-counter therapy or non-responsive/intolerant of the triptans.

Yet again, this study is completely funded by Biohaven Pharmaceuticals, with all the authors employed by and owning stock.

So, the same critiques as yesterday apply – and it’s transparently just another page out of the typical pharmaceutical corporation playbook: do enough to obtain approval, and then let the marketing operation kick into high gear.

“Efficacy, safety, and tolerability of rimegepant orally disintegrating tablet for the acute treatment of migraine: a randomised, phase 3, double-blind, placebo-controlled trial”
https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(19)31606-X/fulltext

Addendum: The protocol does have a non-redacted tidbit of a comparison between rimegepant and sumitriptan, based off some unpublished data from a Phase 2b trial. To be taken with a grain of salt, to be sure, but obviously this would not have looked good for rimegepant if any sort of active comparator would have been used:

The New Cutting-Edge Treatment for Migraines

The new sexy business – by which I mean “profitable” – in the treatment of chronic migraine is therapy targeted at the calcitonin gene–related peptide receptor. The past few years have brought several of these to market as ongoing maintenance therapy. This looks at a different application – acute migraine. Their proposed unmet need – the slice of patients for whom various triptans are ineffective.

This is a multi-center, randomized, double-blinded, placebo-controlled trial of rimegepant, an orally-administered CGRP antagonist, administered as a single-dose for acute migraine headache of moderate-to-severe intensity. These authors randomized 1,186 patients in a 1:1 ratio, with the primary end point being freedom from pain at 2 hours after the initial dose.

The winner: rimegepant.

Yes, rimegepant is superior to doing nothing at all for your migraine headache.

And just barely – 19.6% response to rimegepant, compared with 12.0% response to placebo.

The clinical value here is virtually negligible. And, helpfully, the authors provided the primary limitation of this trial for you in the text:

“First, the trial did not include an active comparator to rimegepant.”

The authors note in their introduction many patients receiving triptans do not have a response – 34%! This, however, implies 60+% of patients do have a response – far superior to rimegepant. Then, patients also have an array of over-the-counter options including acetaminophen, ibuprofen, and various caffeine-containing combination therapies. Treating moderate-to-severe migraine attacks with placebo borders on – if not crosses into – unethical territory. Even if this therapy didn’t have a dismal response rate, we still need to generalize it one step further to those who are non-responders to triptans to even have an indication – which would then be the proper enrollment criterion for a trial to ensure the same physiologic features making a patient a triptan non-responder weren’t also a CGPR antagonist non-responder.

Now, no one opposes further exploration of alternative, effective treatments for migraine – headaches, particularly migraine headaches, are relatively common presenting complaints in the Emergency Department. While we have effective abortive treatments for such, there is tremendous value in having a wider array of options for use at home, considering the direct and indirect resource costs and human suffering associated with headaches requiring Emergency Department evaluation and treatment.

But this is junk science – an advertorial in a journal continually proving itself to have virtually no worthy editorial standard:

“Biohaven Pharmaceuticals sponsored the trial, supplied the trial agents, reviewed the trial design, collected the data, and performed data management and analysis. The manuscript was written with the assistance of a medical writer funded by Biohaven Pharmaceuticals. All the authors have confidentiality agreements with Biohaven Pharmaceuticals, either as a condition of employment or in their role as consultants.”

Yet another utter embarrassment for the New England Journal of Medicine.

“Rimegepant, an Oral Calcitonin Gene–Related Peptide Receptor Antagonist, for Migraine”

https://www.nejm.org/doi/full/10.1056/NEJMoa1811090

Addendum: There’s a second study, as well, published in The Lancet with similar results. And, furthermore, deep in a non-redacted part of the protocol, the authors share these unpublished Phase 2b trial results. To be taken with a grain of salt, to be sure, but obviously any comparison with an active drug would have looked much worse for rimegepant:

Pediatric C-Spine Injury Risk Factors

It would seem the Pediatric Emergency Care Applied Research Network (PECARN) is gearing up to develop another decision instrument – this time for cervical spine injuries.

This is a prospective, observational study of 4,091 pediatric blunt trauma patients across four pediatric level-1 trauma centers, surveying treating providers about the presence or absence of factors suspected to be implicated with cervical spine injuries. The factors were selected based on previous studies, as well as those suspected as having potential physiologic plausibility and good interrater reliability. The stated purpose – to ultimately develop a decision instrument akin to their prior work for clinically important minor head injury.

Overall, the prevalence of a cervical spine injury – vertebral fractures, ligamentous injury, intraspinal hemorrhage, or spinal cord injury – was 1.8%. The vast majority of patients in their cohort (78.2%) underwent some sort of imaging, although only 15.8% underwent CT. The most predictive items identified are those already typically considered: diving injuries, axial loading injuries, clotheslining, loss of consciousness (including intubation), neck pain, altered mental status (frequently associated with obvious head injuries), limited range of motion, focal neurologic deficits, and substantial torso and thoracic injuries. Of the 74 patients with CSI in their cohort, effectively only one would have been missed by a decision instrument based on these factors – a fall from 10 feet whose symptoms localized to the thoracic spine, and had a C7 burst with T2-T4 compression fractures. Obviously, this was not missed clinically – again revealing the role of clinical judgment outside of any decision instrument.

The most interesting tidbit, leading into the most substantial implications for generalizability, is their note regarding “high-risk MVC”. They comment previous case-control studies determined both predisposing conditions (e.g., congenital abnormalities of the cervical spine) and high-risk MVC were identified as risk factors, whereas in this study they were not. They discuss the low prevalence in their cohort of those with predisposing conditions, and, conversely, the high prevalence of high-risk mechanisms, to justify their lack of multivariate effect on predicting CSI. Even though they did not fall out as predictive elements in this cohort, a future prediction model intended for general use may yet include such features. As such, these data ought not be fully relied upon to downgrade those potential risk factors.

“Cervical Spine Injury Risk Factors in Children With Blunt Trauma”
https://pediatrics.aappublications.org/content/early/2019/06/18/peds.2018-3221

Ye Big High-Sensitivity Troponin Evaluation

After many years of various studies of moderate size looking at the diagnostic performance of high-sensitivity troponin assays, now we have a new entry: the Calculation of Myocardial Infarction Risk Probabilities to Manage Patients with Suspicion of Myocardial Infarction (COMPASS-MI) project.

This is not new data, but rather the power team of Roche and Abbott pooling the results of 15 prior studies to describe the diagnostic performance of their Elecsys and Architect high-sensitivity platforms. Then, there are really two parts of this article. There is an initial analysis looking at the performance characteristics of differing combinations of initial and serial sampling of each. After that, these authors pull in several age- and comorbidity-matched comparison populations and describe the long-term 1- and 2- year outcomes of patients with differing levels of troponin concentrations.

The main product of their work, however, boils down to a set of mildly confusing wheels of data regarding the negative predictive value of various combinations of initial troponin level and serial troponin change, divided up based on whether repeat sampling was performed early or late. These cut-offs are further divided on the wheel regarding the proportion of the population with a certain risk level for 30-ay MI or death.

The end result, combined with the various massive supplementary appendicies, are massive amounts of data to help systems using these assays tailor their practice patterns to their desired level of sensitivity and specificity. The authors are not specifically prescriptive in any one cut-off, but rather try to provide as much data as possible. Prevalence of nSTEMI in their population was about 14%, meaning the negative predictive values are only generalizable to to similar patient demographics.

If you’re using these assays, this is quite important work to help assist in interpretation. If not, considering there’s no comparative data to conventional assays, it seems to have limited utility.

It should finally be noted virtually all the listed authors of this work receive some financial support from the manufacturers of these assays.

“Application of High-Sensitivity Troponin in Suspected Myocardial Infarction”
https://www.nejm.org/doi/full/10.1056/NEJMoa1803377

Yet Another Febrile Infant Rule

The Holy Grail in the evaluation of infants of less than 60 days remains safe discharge without a lumbar puncture. Boston, Philadelphia, Rochester, Step-by-Step and others have tried to achieve this noble goal over the years. And now, the Febrile Young Infant Research Collaborative has tossed their hat into the ring.

In this retrospective query of their Pediatric Health Information System and other electronic medical records, these authors identified 181 non-ill appearing patients across 11 Emergency Departments with invasive bacterial infection, defined as bacteremia in either blood or cerebrospinal fluid. Using 362 matched controls as a comparison cohort, these authors used the typical logistic regression route to tease out the strongest predictors of IBI – age in days, observed temperature, absolute neutrophil count, and urinalysis result. Subsequently, they condensed the continuous variables into cut-offs maximizing area under the curve. These cut-offs were then incorporated into a scoring system based on the strength of their adjusted odds ratio, and then the final output was validated on the derivation set using k-fold cross-validation with 10 sets.

The final result using their best cumulative score cut-off: sensitivity of 98.8% (95% CI 95.7-99.9) with 31.3% specificity. The two cases missed were that of a 3-day old and a 40-day old otherwise afebrile in the ED with normal UA and an ANC <5185. The authors ultimately conclude their score, if validated, may have best value as a one-way prediction tool primarily to reduce current routine invasive testing, owing to its poor specificity. Certainly, I agree it does not have much value in those who might otherwise not undergo testing; a more specific risk score may be better, if not clinician gestalt.

The other tidbit I might mention is whether there could be value in incorporating time-of-onset of fever into their evaluation. We’ve seen in other studies a few of the fallouts with regard to sensitivity of IBI stem from recency of illness onset, and it may be falsely reassuring to find a normal ANC early in an illness course. Furthermore, these authors do not specifically mention whether the lack of fever in the ED could have been associated with prehospital antipyretic use. Finally, their data collection does not appear to incorporate respiratory swab results; readily available respiratory viral panel results may also prove useful in ruling out IBI.

While these data are certainly alluring, considering the desire to avoid invasive procedures in young infants, substantial prospective work is still likely required.

As a sad aside, the authors state:

However, these criteria were developed >25 years ago, and the epidemiology of serious bacterial infections has changed considerably since that time.

Unfortunately, as vaccination frequency continues to decline, even since patients were enrolled for this study, our “modern” cohort may better begin to resemble that of 25 years ago.

“A Prediction Model to Identify Febrile Infants ≤60 Days at Low Risk of
Invasive Bacterial Infection”

https://www.ncbi.nlm.nih.gov/pubmed/31167938

Upping Your CSF Game

WBCs? Glucose? Gram stain? Next-generation genetic sequencing?

It’s the NEJM again, so you know the fingerprints of financial and professional conflict of interest pervade, but this study is still fairly typical of the types of infectious disease diagnostics on the horizon. Why wait for any specific organism to grow – over the course of days – when you can simply try and match DNA fragments floating around to those of various viral and bacterial pathogens?

The promise probably doesn’t quite meet the hype in this study, based out of UCSF, where many of those working on the project hold shares of the patent on the technology. In this prospective multicenter study, these authors recruited patients, ostensibly, who were diagnostic challenges – “idiopathic meningitis, encephalitis, or myelitis in patients who had not received a diagnosis at the time of enrollment”. The vast majority of those enrolled were ultimately encephalitis and meningitis. Then, this wasn’t specifically a formal trial as much as it was a collected case series with a 1-year convenience time frame, constrained by funding and testing capacity.

The authors screened 482 patients for a final study population of 204. Of these 204, their next-generation sequencing methods made a diagnosis in 32. Of these 32, 19 had already been made by further directed clinical evaluation. Of those final 13, then, in which the NGS assay was the only method of diagnosis, this information augmented clinical management in 7. The supplementary appendix details these specific impacts on management – although, in reality, few of the vignettes are terribly compelling. A handful of cases confirmed a suspected diagnosis, leading to clinicians to narrow antibiotic or antifungal therapy, while others “reassured” clinicians they were on the right course. The NGS assay did, however, occasionally detect clinically important pathogens and guide directed treatment, including Nocardia and S. mitis meningitis whose conventional testing was otherwise negative. Unfortunately, despite the addition of this testing, no conclusive final diagnosis was ever made in half their cohort.

At present, this sort of testing is not likely to be within the scope of the Emergency Department – these represent complex cases with low diagnostic yield, and even while this method picks up some new diagnoses, it also misses others established by conventional means. That said, this sort of technology will likely yet only improve, decrease in cost, and additional applications will edge closer to mainstream care.

“Clinical Metagenomic Sequencing for Diagnosis of Meningitis and Encephalitis”

https://www.nejm.org/doi/full/10.1056/NEJMoa1803396