More public health doom and gloom from the infectious disease world – multiple bacterial found carrying stable and transmissible plasmids for the NDM-1 gene, a form of beta-lactamase that endows bacteria with carbapenem resistance. In the sewage. And in the drinking water.
These are two articles coming to us from January’s NEJM – one from Pittsburgh and one from Finland. For a ridiculously prevalent disease managed daily in the ambulatory setting by thousands of providers across the world, we really don’t know what we’re doing. Firstly, we know there are bacteria in the middle ear – plenty of studies have aspirated out a variety of pathogens. Secondly, undertreated AOM may lead to suppurative complications – the surgical emergency of mastoiditis. Finally, however, we also know the natural history of AOM is to resolve without intervention and treatment in most cases.
The commentary accompanying these articles seemed to think these two studies tilted the balance in favor of routine antibiotic use for AOM. And, you can read the articles in that fashion – the Finnish article makes an argument that by their standard of care, the placebo group had 45% treatment failure and required antibiotic rescue 30% of the time. However, I read the article and their definitions of treatment failure are, for the most part, not clinically relevant. A red ear that still looks red on day three does not predict much – and, actually, a lot of their primary outcome measures were not statistically significant until they added in the 30% they used rescue treatment on to boost their definition of treatment failure. So, this study doesn’t tell much much – except that Augmentin causes diarrhea 40% of the time.
The Pittsburgh study reads much more straightforward and, partially, justifies my criticism of the Finnish study. By day 7, 80% of their Augmentin group had symptom improvement compared to 74% of their placebo group. However, the ears still looked terrible in the placebo group – but, obviously, weren’t correlating with symptoms. So, is it clinically relevant to define treatment failure based on the appearance of the ear? I would say that improvement in symptoms is the appropriate measure of clinical cure – and in that respect, the 80% vs 74% numbers match up better with previously published literature.
To me, these articles don’t help. I don’t necessarily mind the AAP recommendations of treating AOM with antibiotics under a year of age – although, really, after their third set of immunizations, it’s not as though this is going to be a nidus for SBI. Clearly, antibiotics offer some advantage – but not to everyone. When I have a three year-old I have to hogtie to get a one-second glimpse at a red TM, how do I know whether this is a patient whose AOM will resolve on its own, or whether he falls into the subset of patients who will derive benefit from antibiotics? And, is that 6% absolute difference in clinical outcome worth giving 40% of children horrible diarrhea (not that anyone uses Augmentin first-line for AOM, anyway). With the NNT with antibiotics to prevent one case of mastoiditis estimated at 4100 from cohort data from the United Kingdom, that’s a lot of useless (and harmful) antibiotic prescriptions. If we want to reduce the amount of antibiotic resistance in this country, I think these studies support a conservative non-antibiotic strategy initially in appropriately selected patients.
Addendum 5/11: Scott Weingart & David Newman did a bit of a rant on this topic on the most recent EM:RAP where they attacked the Pittsburgh article specifically regarding result reporting integrity. Dr. Newman delved into far more detail regarding the multiple primary outcomes listed and found the original research protocol listed only time to resolution of symptoms as the primary outcome – which was statistically equivalent between the two groups. He and Dr. Weingart seem to suggest there was not sufficient editorial oversight regarding the publication of this article due to these flaws, and that the author’s conclusions are suspect to the point of disingenuous.
Discussion today in the public forum about this article:
This is a fun study because it’s always nice to have new things to try for common problems. Keeps life interesting.
I started out residency taught to use metoclopromide+diphenhydramine for treatment of refractory headache in the Emergency Department. And then I discovered droperidol. Yes, there are studies out there that say prochlorperazine is equivalent to droperidol in efficacy, but prochlorperazine gives people the same akathisia that metoclopromide does. Droperidol kills people dead, if you believe the black box – though I don’t. The QT-prolongation is essentially no different than ondansetron, the supposedly-safe alternative we now use for nausea.
In any event, now you can add olanzapine to your mix. You can legitimately critique the study because the p-value for pain improvement between olanzapine and droperidol was actually 0.30 in favor of droperidol – so without more power and/or a second confirmatory study, you can say it really might not be as effective. But, the good thing is, nearly everything has some legitimate effectiveness – and the more different classes of medication you have available to knock down that headache, the better.
Low-risk chest pain – if your ED doesn’t already have a chest pain unit set up for you to painlessly move patients through their enzymatic and non-invasive testing, you’re probably trying to find safe ways to discharge your chest pain patients home to avoid the repetitive calls to an unsympathetic hospitalist. Problem is, without some kind of imaging or functional study, you’re going to invariably get burned. This is another one of the TIMI-score-plus-X attempts at risk-stratifying patients in a prospectively applied dry run of their protocol. It’s TIMI 0 patients plus normal EKG plus negative zero and two-hour CKMB/troponin/myoglobin. Basically, 10% of their chest pain cohort fit this essentially zero-risk profile and were enzymatically ruled out. And 0.9% (0.02 to 2.1%) of this slam-dunk non-cardiac group came back with an MI within 30 days.
Now, for a rational person who thinks that we’re spending altogether too much money and resources to capture every last potential cardiac event – that sounds pretty reasonable. Home with follow-up. The problem is, the non-invasive testing in basically the same sort of low-risk cohort, whether stress or CTA, the negative tests have 6+ month event-free periods. So, the standard of care is unfortunately moving away from “no heart attack today!” to prognosticating distant events.
The other great thing about this article was their mini systematic review where they say there’s 115 of these prediction rules in the literature in the last fifteen years. Clearly someone everyone wants, but also something we can’t get right….
These authors show us something very interesting about “negative” troponin testing in their prospective review. Like my facility, they had an assay with a certain level of detectability. I will say, though, that their assay detection level of 0.20 ng/mL is different than our assay, that goes down to 0.02 ng/mL. But, basically, they implemented a new assay that was sensitive down to 0.05 ng/mL but didn’t tell their clinicians any of the newly detectable troponin numbers below 0.20, but they built a database of the cohort in that detectable-but-previously-undetectable range. Then, they showed clinicians their new detectable numbers and let them do what they wished with it. It turns out, the population with the 0.05 to 0.20 troponins had similar cardiac morbidity/mortality risk as their >0.20 troponin population – while the truly undetectable did much better over their 1-year follow-up.
Additionally, after the new threshold was introduced, clinicians treated the 0.05 to 0.20 more similarly to the >0.20 group with additional testing, cardiology consultation, and medications – and their outcomes improved 20%.
So, this study really makes you think what it means to have a “negative” troponin and wonder what the disposition and follow-up should be at different points. At my institution, we’ve traditionally said 0.10 ng/mL is the cut-off for “positive”, and the lesser, yet still detectable troponins, are “negative”. But, this study clearly suggests that a lot of the patients we’re labeling “negative” really need to have more aggressive treatment. This study doesn’t tell us whether they need to be inpatient, or have expedited follow-up, or precisely how to manage them in order to improve their outcomes, but it suggests a lot of great new questions to ask regarding how to use the information we receive from our troponin assay.
I was really hoping this would be a great article that convinced me that my hesitancy towards cardiac CTA is unfounded. I feel, based on the literature, that we’re misusing cardiac CTA – or at least, the current generation of technology and reconstruction methods aren’t leading us in the right direction. Angiography, whether radiographic or invasive, describes anatomy, and then we use the anatomy as our basis whether to attribute chest pain to cardiac causes or not. Many situations, this works – the STEMI goes to the cath lab and the occlusion correlates with symptoms. But, we’re trying to use CTA in our low-risk population to draw conclusions about the etiology of chest pain – and it’s much harder to say someone’s troponin-negative and EKG non-specific chest pain comes from a stenosis of a certain percentage.
The problem with their article is that they completely underestimate the number of false-positives cardiac CTA is generating. There are several articles out there showing that the population considered for cardiac CTA is generally a population that just does great in follow-up, and that the number of negative follow-up studies generated after cardiac CTA – nuclear stress and invasive angiography – tend to far outnumber the number of positives. They base their cost estimates on numbers that just don’t reflect reality, and I just can’t believe that cardiac CTA is a test that saves money and gives me better answers compared to functional cardiac testing. If you wanted to use it as a screening tool for identifying a population that needs aggressive secondary-prevention of progressive atherosclerotic coronary disease, it would do great at that – but we know that plenty of ACS comes from ruptured plaque and hemodynamically insignificant disease, and someone who had a “negative” cardiac CTA that morning doesn’t preclude them from needing an enzymatic rule-out.
Here’s an article with a lot of great numbers to keep you from sleeping well at night. It’s a prospective look at the symptoms patients present with when their eventual diagnosis is STEMI or NSTEMI. 6.4% of their STEMI patients and 5.6% of their NSTEMI patients didn’t complain of any chest symptoms (pain, pressure, etc.). There’s a lot of arm pain, epigastric pain, shoulder pain, and then multiple anginal-equivalents on their list of symptoms, but it’s a great example showing that if you write off ACS in your patients without chest pain, you’re going to get burned – 1 in 16 STEMIs in their cohort. Even better, only 45.1%/53.9% of their patients thought their symptoms were cardiac related, which probably means there’s a population of STEMIs just sitting at home figuring they’ll feel better in the morning….