Levetiracetam vs. Phenytoin

The epic, classic showdown from time immemorial: new vs. old.

But, more specifically, these are two trials set to determine relative utility of each in pediatric seizures, vying for the coveted “second-line” therapy recommendation once benzodiazepines have failed. The thought and hope is, of course, the newer agent – levetiracetam – is at least as efficatious, if not moreso, as it can be infused more quickly. The authors then propose levetiracetam is associated with fewer long-term adverse effects and medication interactions during oral maintenance, and, as such, allows for convenient continuation after intravenous initiation.

Generally speaking, these two trials are very similar – both open-label trials randomizing pediatric patients with ongoing seizures, both analyzing about 250 patients … and both demonstrating effectively similar results by different routes. EcLiPSE, interestingly, was designed as a superiority trial, with a primary outcome of time from randomization to seizure cessation. Median time to seizure cessation not statistically different at 35 minutes for fosphenytoin and 45 minutes for levetiracetam, with similar numbers of treatment failures for additional anticonvulsants or intubation. In ConSEPT, the primary outcome was cessation of seizure activity within 5 minutes of the end of the infusion, and here results favored phenytoin at 60% versus 50% for levetiracetam.

Effectively, however, the sample sizes are small enough, the types of patients heterogenous enough, and the differences small enough, the Bayesian interpretation is probably a wash. These are both fine second-line options, but these trials do not provide any data supporting levetiracetam as a superior option.

“Levetiracetam versus phenytoin for second-line treatment of paediatric convulsive status epilepticus (EcLiPSE): a multicentre, open-label, randomised trial”
https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(19)30724-X/fulltext

“Levetiracetam versus phenytoin for second-line treatment of convulsive status epilepticus in children (ConSEPT): an open-label, multicentre, randomised controlled trial”
https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(19)30722-6/fulltext

Back Pain is a Pain

The authors’ bottom line:

“Adding baclofen, metaxalone, or tizanidine to ibuprofen does not appear to improve functioning or pain any more than placebo plus ibuprofen by 1 week after an ED visit for acute low back pain.”

There were 320 patients with atraumatic back pain randomized to one of four arms in a 1:1:1:1 manner at two urban Emergency Departments, with follow-up one week later. As noted above, the differences in improvement on the Roland-Morris Disability Questionnaire were similarly scattershot across groups. The waterfall graph probably shows this best:

Notes:

  • The prescribed dose of ibuprofen was only 600mg every eight hours. I’m more a 500mg naproxen twice daily sort of guy, but this wouldn’t necessarily obscure differences between groups.
  • Half of patients continued to complain of back pain at seven days, including a third for whom it was moderate or severe.
  • Adverse events were rare in all groups.

Pretty dismal and disappointing, as so many of our patients in the ED are coming in because they’re already taking some over-the-counter analgesic without perceived relief. There is clearly an unmet need for something to fill the void, hence our attempted use of all these unproven adjuncts. Unfortunately, these folks are stuck – it’s going to be 2-3 days until they will be able to return to their usual activities, if not more, and none of these are adding any value.

“A Randomized, Placebo-Controlled Trial of Ibuprofen Plus Metaxalone, Tizanidine, or Baclofen for Acute Low Back Pain”

https://www.ncbi.nlm.nih.gov/pubmed/30955985

A Brief BRUE Follow-Up

The “apparent life-threatening event” has long since been replaced by the “brief resolved unexplained event – an event occurring in an infant <1 year of age with:

  • Cyanosis or pallor
  • Absent, decreased or irregular breathing
  • Marked changes in tone
  • Altered level of responsiveness.

This is just a brief, single-center, retrospective chart review identifying children hospitalized for BRUE and their outcomes up to 5 years. These authors identified 120 hospitalized infants under one year of age meeting criteria for BRUE, and performed telephone follow-up at least 6 months after the event. Most children hospitalized were less than 1 month of age, and about half were hospitalized for apnea or breathing issues. Of the 87 they were able to contact, none had died or developed chronic medical illness, 71 had developed normally, and the remainder developed either a global or verbal developmental delay.

This is a small sample, to be sure, but these data suggest children hospitalized for BRUE are not specifically at risk for long-term poor neurologic or cardiac outcomes above the baseline population level.

“Long-Term Follow-Up of Infants After a Brief Resolved Unexplained Event–Related Hospitalization”

https://www.ncbi.nlm.nih.gov/pubmed/30951030

A (Mostly) Unnecessary Antivenin

We’re something of experts in spider bites in the Emergency Department. Black widow-type spider bites, however, are a little more exceptionally interesting than the average patient encounter. Latrodectism is usually nonfatal and easily managed with conservative treatment in the ED, although an immunoglobulin-based antivenom exists. Adverse effects from this antivenom are likely high-risk than any symptomatic benefit.

Now, a new antibody-fragment antivenom is vying for space on your formulary. In this study, 64 patients with clinical lactrodecism and significant baseline pain were randomized to either up to two doses of antivenin thirty minutes apart, or matching saline placebo. The primary outcome was “treatment failure”, a composite endpoint of failure to sufficiently decrease pain on the VAS scale, provision of a prescription analgesic, or a dose of the currently available antivenin.

Overall, the intervention met their predefined primary outcome – although, interestingly, half the antivenin cohort still had “treatment failure”. Then, almost a quarter of patients in the placebo cohort achieved full recovery. This remaining sliver of patients represents the measured effect size of the intervention – but the question remains the value and clinical relevance of their composite outcome. Mortality from lactrodectism is vanishingly low, and this intervention is not touted as being protective. Then, if the concern voiced by the authors is of short-term morbidity and ED recidivism in those whose pain is not adequately controlled without antivenom, then their primary, or at least a secondary outcome, ought to address this. Finally, the sample size is too small to draw any conclusion regarding safety.

This will undoubtedly be expensive and may likely expire in your hospital pharmacy prior to use. This trial gives us a small amount of information regarding its utility, but there’s little clear value demonstrated here.

“The Efficacy of Antivenin Latrodectus (Black Widow) Equine Immune F(abʹ)2 Versus Placebo in the Treatment of Latrodectism: A Randomized, Double-Blind, Placebo-Controlled, Clinical Trial”

https://www.sciencedirect.com/science/article/pii/S019606441930109X

More Cardiac Stress Test Futility

Once upon a time, it was believed important to perform cardiac stress testing in patients with chest pain and potential acute coronary syndrome. Intuitively, this makes little sense – a stress test may identify obstructive coronary disease, but this is rarely the culprit for acute coronary syndrome, and certainly only rarely the cause of chest symptoms in a patient at low-risk for coronary artery disease. Unfortunately, the American College of Cardiology/American Heart Association have had a nonsensical recommendation for noninvasive testing within 72 hours of an index visit on the books for quite some time – leading, to put it mildly, to a test or two.

This observational data set evaluates the outcomes of patients in the Kaiser Southern California system who were referred for outpatient stress testing following an encounter in the Emergency Department. They tracked 7,988 patients for a month after their ED encounter, 2,497 of whom underwent stress testing within 3 days, 4,695 within 4 to 30 days, and 796 who never showed up for their referred testing. Most stress tests were exercise or pharmacologic stress ECG, with the minority stress echocardiograms or myocardial perfusion imaging.

Patients undergoing testing were not devoid of risk factors: an average age of 55 years, most were overweight, and over half had at least one risk factor for coronary artery disease. Within 30 days, fewer than 1% were diagnosed with an acute MI, and a handful of those underwent PCI or CABG. There were tiny differences between groups, as none of the patients who skipped their stress test underwent subsequent revascularization. There were no deaths in any cohort.

The narrow view here in this article is there is no apparent benefit to undergoing stress testing within 72 hours as compared with a longer timeframe. The wider view is yet another piece of information showing the general disutility of stress testing. These are not randomized cohorts, nor is 30-days a long enough window to detect any potential benefits to the stress test – as measured by decreased morbidity or mortality as relating to a timelier PCI or CABG. However, even these interventions were rare enough the effect size from any benefit is bound to be so small as to represent low-value care. It is absolutely reasonable to suggest the Bayesian pendulum for the valuation of stress testing is swinging the other direction – and those who advocate for stress tests ought to generate data to support its targeted use, rather than for the opposition to generate data contrary to its assumed routine utility.

“Evaluation of Outpatient Cardiac Stress Testing After Emergency Department Encounters for Suspected Acute Coronary Syndrome”

https://www.sciencedirect.com/science/article/pii/S019606441930054X

Any Troponin is Bad Troponin – Gender-Specific Edition

High-sensitivity troponins mean a lower limit of detection. Picking up these lower quantitative values for circulating troponin – and new reference limits for the 99th percentile of normal – has required an adjustment in perspective with respect towards making the diagnosis of acute coronary syndrome. Now, the question with these more sensitive assays becomes: should we adjust our clinical considerations to incorporate sex-specific reference intervals?

This brief analysis from the UTROPIA study looks specifically at the downstream MACE in patients whose serial troponin measurements fall between the limit of detection and the sex-specific 99th percentile intervals. For this Abbott assay, that means 34 ng/L for men and 16 ng/L for women. In their 180-day follow-up period, they found the incidence of major adverse cardiac events was vanishingly small for those with undetectable levels of circulating troponin. However, those with any circulating troponin – even below the 99th percentile reference interval – were vastly more likely to experience an event within the next 180 days, reaching about 10% incidence of MACE. Importantly, however, the distributions of probability for downstream MACE were similar with regard to the measured value with respect to the sex-specific 99th percentile.

This confirms some of what we already knew: any troponin is bad troponin, even if it’s lower than the 99th percentile. Then, this also validates sex-specific 99th percentiles, as percentile levels conveyed similar risk between men and women.

Just another insight into the next level of sophistication for use of these assays in assessing patients with potential ACS, and for downstream anatomic assessment and preventive interventions.

“Clinical Features and Outcomes of Emergency Department Patients With High- Sensitivity Cardiac Troponin I Concentrations Within Sex-Specific Reference Intervals”

https://www.ahajournals.org/doi/10.1161/CIRCULATIONAHA.118.038284

Lacunar Infarcts & Thrombolysis

For some period of time, folks have debated the utility of thrombolysis in lacunar infarcts. The underlying concern is with regard to their underlying suspected pathology relating to non-thrombotic occlusion of small perforating arteries, in contrast to the process seen in small- or large-vessel stroke. This little subgroup analysis of WAKE-UP – the MRI-driven tissue-based trial of alteplase for ischemic stroke – tries to shed further light on this specific concern.

Of the 503 patients included in WAKE-UP (out of 1,362 patients screened), 108 had imaging-defined lacunar infarcts. The median NIHSS of these patients was 4 to 5, and about half were randomized to alteplase and half to placebo. Overall, this subgroup – underpowered for any definitive conclusion – demonstrated similar outcomes as those whose stroke subtype was not lacunar.

The issue is not so much the finding observed here, but the effort in the Discussion and accompanying editorial to generalize WAKE-UP to all strokes. There is only a loose association between DWI and FLAIR findings and predicting time of stroke onset in their cited reference. A little fewer than 2/3rds of strokes of ≤4.5h age seem to have positive DWI and negative FLAIR, and this study enrolled only a tiny fraction of patients with potential lacunar infarcts.

Long story short, a treatment effect observed in this tissue-based enrollment cohort cannot reliably predict treatment response for lacunar strokes screened and treated based on routine non-contrast imaging. Most patients screened for WAKE-UP were excluded based on not meeting imaging criteria, potentially around half of whom were otherwise within 4.5 hour stroke onset (based on their citation above). Thrombolysis does benefit the patients in WAKE-UP, overall, but this almost certainly represents the ceiling for a positive effect size – and in routine practice, effectiveness is likely much lower.

Related aside: when we start routinely screening strokes with MRI in that happy future time, do we exclude DWI+/FLAIR+ from thrombolysis, even if within the “treatment window”? I would think so.

“Functional Outcome of Intravenous Thrombolysis in Patients With Lacunar Infarcts in the WAKE-UP Trial”
https://jamanetwork.com/journals/jamaneurology/fullarticle/2729091

The “True” 99th Percentile for Troponin

Not. All. Troponin. Elevations. Are. Acute Coronary Syndrome.

This is just a simple, prospective, observtional study of 20,000 high-sensitivity troponin I results from a single hospital in the United Kingdom. All blood samples submitted to the lab were evaluated with a troponin assay, regardless of clinical indications or reasons for order. About half of included samples were outpatients, a quarter were Emergency Department patients, and the remainder were inpatients.

The manufacturer-stated upper limit of normal for the assay – Beckman Coulter Access AccuTnI+3, specifically – is 40 ng/L, with a coefficient of variation of less than 10% at that cut-off. Outpatients were the least likely to exceed this limit, at 2%, with an observed 99th percentile of 65 ng/L. In the Emergency Department, 6.1% exceeded the manufacturer-defined cut-off, with an observed 99th percentile of 215 ng/L. Finally, 7.3% of inpatients exceeded the manufacturer cut-off, with an observed 99th percentile of 563 ng/L.

The authors also tracked whether the patients had a troponin ordered by the clinical team, or had a final diagnosis of acute myocardial infarction, and perform some multivariate analyses with these patients excluded. Those only comprised 1,829 patients from this cohort, and excluding these patients made only small impacts upon their new “normal” 99th percentiles. Effectively, the point of this article is a reminder of the effect of concomitant illness on circulating troponin levels – and to take into account the likelihood of an ACS when interpreting troponin elevations above the manufacturer’s 99th percentile cut-off.

“True 99th centile of high sensitivity cardiac troponin for hospital patients: prospective, observational cohort study”
https://www.bmj.com/content/364/bmj.l729

Emergencies in Medicine

Just a quick note to follow-up this splendid little conference in Park City, UT, where I featured in a debate regarding the utility of tPA for stroke. We had a generally respectful discussion about the state of stroke care and our vision for the future.

Unfortunately, it – as you might expect – looks a lot like a future where tPA flows like water.

The overall gist of the presentation by Chris Lewandowski, one of the original NINDS investigators: pool the trial data, and the benefit is clear. My gripes, well-chronicled on this site: benefit is not uniformly distributed, we should tailor its use, rather than expand it.

Interesting tidbits from the follow-up discussion:

  • No one is going to re-do NINDS in this country. They couldn’t even complete PRISMS because too many mild, non-disabling strokes were already being treated.
  • Treating vast numbers of stroke mimics is not troublesome to them – Lewandowski claims to have never heard of one ever having a bleed.
  • The expected benefit to mobile stroke units relates to “fresh” clot being more likely to lyse, as much as the brief time savings.
  • Studies like TIMELESS and other tissue-based thrombolytic studies will likely extend the treatment window, just like WAKE-UP.
  • No qualms about treating NIHSS scores of 1 causing only mild disability (say, unilateral leg weakness). They’ve seen that some NIHSS 1 deteriorate, and believe tPA will prevent it.
  • They are utterly comfortable with forever using NINDS as their default NNT/NNH consent – and no problem using the same numbers now matter what the NIHSS. Except, they just reduce their risk number for harm, as ICH is generally related to infarct size/NIHSS. This drives me mad.
  • They feel strongly tPA has been proven “cost-effective”, while I would note those analyses are based on assumptions and models not matching current practice and treatment population.
  • Treating cervical artery dissection with tPA is favored to them because they expect the clot that showers distally will benefit from lysis, though they agree there is no evidence to support their claims.
  • They agreed that, overall, the recommendations issued by the AHA overstate the strength of the evidence.

Would be fun to do again – even if it is effectively just shouting into the wind!

PE in Pregnancy & YEARS Protocol

So, I generally like the YEARS protocol. It’s an incremental step forwards towards incorporating a pretest likelihood of disease into the interpretation of the D-dimer result. At the least, considering our reckless overdiagnosis of pulmonary embolism, it doesn’t make the situation worse.

Now, pregnancy. There is a mild increase in risk for PE while pregnant, and subsequent puerperal risk is even higher. Unfortunately this leads to a paradox: the index of suspicion for PE during pregnancy is so high, the yield of testing is frighteningly low – on the order of 5% or less for those undergoing evaluation for PE. Little prospective data regarding safe exclusion of PE during pregnancy are available.

These investigators – the Artemis Study – applied the YEARS algorithm, to the diagnosis of PE during pregnancy, attempting generally to demonstrate its safety while describing its yield and test characteristics. The primary modification to the baseline YEARS algorithm was patients having clinical signs of deep-vein thrombosis underwent compression ultrasonography prior to being evaluated with D-dimer and potential CTPA.

They enrolled 498 patients in whom PE was considered a relevant clinical diagnosis, about half of whom met no “high-risk” YEARS criteria. About a fifth of the patients were first trimester, and the remainder were split between second and third. The news is good and bad, unfortunately. Overall, the incidence of PE in their cohort was only 4.0% – typical of our deranged gestalt for PE in pregnancy. In the first trimester, their algorithm excluded PE without CTPA in 65% of those enrolled – meaning only 26 CTPA were indicated to diagnose the 5 PEs in this cohort. This is a reasonable yield.

However, second and third trimester excluded only 46% and 32% of patients from CTPA – meaning 261 CTPAs would be allowed to diagnosis 15 PEs – a yield of only 5.7%. This is better than performing a CT on everyone, but it’s still abysmal. This results, effectively, from the gradual physiologic increase in D-dimer throughout pregnancy – from a median of 505 ng/mL in their first trimester cohort, to 730 and 1,120 in subsequent trimesters.

So, while their algorithm is clearly safe – only one DVT occurred within their 3-month follow-up period, basically the expected rate of occurrence in their enrolled cohort – it’s not the final answer with regard to pregnancy. The next likely step required is to use our observational data to test specific trimester-related normals and pretest-related multipliers to find the optimal cut-offs, such that second and third trimester performance may approach that of the first.

“Pregnancy-Adapted YEARS Algorithm for Diagnosis of Suspected Pulmonary Embolism”
https://www.nejm.org/doi/full/10.1056/NEJMoa1813865