The content of these recommendations is what made the rounds in various news outlets – the first begrudging revelations that antibiotics are possibly unnecessary for many of acute otitis media. This isn’t news to us, of course, but it’s entertaining to see the precise moment the Rock of Gibraltar starts to make its slow course corrections.
As far as clinical policy statements go, however, this is beautifully constructed. For every actionable statement, these authors offer a concise summary of the benefits, harms, value judgements, intentional vagueness, patient preferences, and exclusions. Whether I agree with the hairs they split on each recommendation is almost overwhelmed by how pleasant it is to understand the basis of their reasoning.
My big irritation: their implication that symptom severity or temperatures greater than 102.2F are somehow specific for bacterial disease more likely to benefit from antibiotics. Odd – or am I missing a notable piece of literature? Please point it out if so!
The other new item of interest is the “Strong Recommendation” for analgesic treatment in cases of AOM. Thank goodness!
“The Diagnosis and Management of Acute Otitis Media”
www.ncbi.nlm.nih.gov/pubmed/23439909
Back For More With Cangrelor
Two negative studies weren’t enough to stop the Medicines Company from CHAMPION PHOENIX, the third attempt at demonstrating cangrelor is useful during PCI.
Cangrelor is a direct-acting platelet adenosine diphosphate inhibitor – same as prasugrel and clopidogrel – that differs in its half-life and route of administration. Rather than clopidogrel, which is a long-acting oral loading dose during STEMI, cangrelor is a continuous intravenous inhibitor that wears off after minutes. This has some theoretical advantages, such as when multiple lesions are found on invasive angiography and coronary bypass need not be delayed for the antiplatelet effects of clopidogrel to wear off.
So, PHOENIX follows up CHAMPION PCI and CHAMPION PLATFORM, each of which were negative for their primary combined endpoint of death, myocardial infarction, or ischemia-driven revascularization. In fact, these studies were stopped at their interim analysis for futility, as they were unlikely to show superiority given the planned enrollment.
So, why does PHOENIX succeed where others have failed? Well, they changed the primary endpoint to a new composite – death, myocardial infarction, ischemia-driven revascularization, or stent thrombosis. And, PHOENIX shows a 0.8% vs 1.4% advantage to cangrelor for stent thrombosis – which accounts for most of the new advantage in primary outcome. In previous CHAMPION studies, stent thrombosis was 0.2% vs. 0.3% and 0.2% vs. 0.6%. So, truly, cangrelor succeeds here mostly because the clopidogrel group fares so much more poorly, rather than on its own merits. Considering 25.7% of the clopidogrel group received only 300mg rather than 600mg, and 36.6% received their clopidogrel during or after PCI, it’s no wonder they had greater stent thrombosis in this study.
It’s pretty clear the Medicines Company learned from its two negative studies and rigged the third one to succeed – and kept it just underpowered enough that severe or moderate bleeding with cangrelor didn’t reach statistical significance (0.6% vs. 0.3% p = 0.09). This reinforces a bias frequently seen in sponsored trials – failure at first begets further trials, while initial success doesn’t lead to confirmatory RCTs that might cast doubts upon the authenticity of the golden goose.
“Effect of Platelet Inhibition with Cangrelor during PCI on Ischemic Events”
www.ncbi.nlm.nih.gov/pubmed/23473369
Dabigatran & CES1 SNP rs2244613
This piece of literature is incredibly important – not because of the specific clinical question it addresses, but, rather, because of the fundamental principle in medicine it illustrates.
In large, heterogenous populations – frequently clinical trials – the risks and benefits are not evenly distributed throughout the cohort. However, when the primary outcome reported, this is based on the aggregate results. This leads us to one of the primary flaws in evidence-based medicine, that statistical power typically comes at the expense of external validity to the individual patient.
This retrospective evaluation of the cohort from RE-LY, the trial comparing dabigatran to warfarin for the prevention of stroke in atrial fibrillation, evaluates genetic polymorphisms and their association with primary and safety outcomes. In short, each minor allele of CES1 SNP rs2244613 was associated with lower trough concentrations and statistically significant interaction between treatment with warfarin and bleeding – hazard ratio 0.59 (0.46 – 0.76) for carriers and hazard ratio 0.96 (0.81 – 1.14) for noncarriers. No difference in the primary stroke prevention outcome was apparent – but outcomes were infrequent and the study was underpowered to detect such a difference.
My take on these results is simple – the 32.1% of patients who are carriers for the rs2244613 allele account for most of the bleeding benefit seen in RE-LY, and the non-carriers have no bleeding advantage compared with warfarin. We ought to be specifically tailoring and reducing the dabigatran treatment population to this subgroup with the most favorable risk profile – and we need to be developing tools that support this kind of individualized treatment throughout medicine.
“Genetic Determinants of Dabigatran Plasma Levels and Their Relation to Bleeding”
www.ncbi.nlm.nih.gov/pubmed/23467860
The Sad Reality of Chest Pain Observations
Chest pain observation units run by the Emergency Department are fairly popular – and it’s easy to see why. It eliminates the need to fight a hospitalist for admission, allows for complete coverage of medicolegal liability, captures another set of billing codes for ED revenue, and keeps the cardiologists happy with a steady stream of interpretation and consultation revenue.
Duke University has one of these such chest pain observation units, and this study is a retrospective evaluation of the subgroup of patients aged less than 40 years. Of the 2,231 patients observed for suspected acute coronary syndrome, 362 met eligibility based on age. Of these 362 patients, median age 36, 238 underwent stress testing and the remainder underwent serial enzymes.
From this cohort, there was a single true positive – defined as a patient who underwent a coronary angiogram with an intervention performed.
There were, however, 14 false positives – indeterminate or positive stress tests and one set of positive biomarkers, leading to five negative invasive coronary angiograms.
The authors sum it up quite nicely: “The extremely risk- adverse physician cannot totally exclude the possibility of ACS based on age, but it seems that routine observation for such patients may cause the potential for as much harm as good.”
“Utility of Observation Units For Young Emergency Department Chest Pain Patients”
www.ncbi.nlm.nih.gov/pubmed/22975283
“ePlacebo”-Controlled Trials?
This is a bit of a fascinating application of clinical informatics – using retrospective patient cohorts and propensity matching techniques to reduce the need for placebo groups in future trials.
This is work done by Pfizer on their own internal database, to address the ethical and financial concerns regarding recruiting large populations for new clinical trials. For example, if you’re testing a new diabetes medication – do you really need a new control group, or can you sort of re-use the control group you had from the previous trial? The answer of course, has traditionally been no – but their answer is yes-and-no. Using their database of over 24,000 trials, they were able to identify 4,075 placebo-controlled groups, with varying degrees of data integrity, crossover, and parallel status. They then suggest these groups could be used, when appropriate, as comparators in future studies in the same domain.
This is certainly an interesting application of clinical informatics – creating temporal databases of clinical trial patients with the potential to augment the evaluation of new medications. What’s nice is that these authors appropriately recognize the limitations of such a database, noting it may only supplement, not replace placebo arms in future trials.
“Creation and implementation of a historical controls database from randomized clinical trials”
www.ncbi.nlm.nih.gov/pubmed/23449762
Stroke or Stroke Mimic? Who Cares!
Suppose you’re “lucky” enough to be taken to an experienced stroke center if you have stroke-like symptoms. After all, they see strokes every day, are experts in the diagnosis of stroke, and have given thousands of patients thrombolytics. However, how often might they be wrong, you ask?
Oh, they estimate about 1 in every 50. But, truthfully, it’s probably much worse.
This is a multi-center observational cohort that purports to identify the percentage of patients treated with tPA and subsequently diagnosed with stroke mimics. Out of the 5518 patients in their cohort, 100 were identified as stroke mimics. Two of the 100 had sICH by NINDS criteria, but none died. Therefore, these authors confirm, tPA is safe even when they’re wrong, and the collateral damage of racing to tPA is low.
Of course, their methodology for identifying a stroke mimic is hugely skewed towards maintaining the diagnosis of ischemic stroke. Only patients in whom clinical details did not suggest a vascular etiology or a clear alternative diagnosis were labeled mimics. Patients with nonspecific features, non-contradictory imaging, or lacking definite evidence favoring stroke mimic remained as diagnoses of acute stroke.
So, even at experienced stroke research institutions – 1 in 50 with the most generous of criteria. What’s the chance real-world performance approaches anything close to this level of diagnostic skill?
The authors, of course, declare multiple financial conflicts of interest with the manufacturer of tPA.
“Safety of Thrombolysis in Stroke Mimics : Results From a Multicenter Cohort Study”
www.ncbi.nlm.nih.gov/pubmed/23444310
Does Size Matter? (Chest Tube Size)
That is, apparently, the question being asked by the trauma folks at Los Angeles County Hospital. Traditionally, traumatic pneumothorax with accompanying hemothorax is routinely treated by the largest chest tube possible. Theoretically, smaller chest tubes will clog with debris or blood clot, requiring additional thoracostomy tubes or interventions.
However, these authors note several simulations of chest tube drainage indicating tubes as small as 14 French may be adequate. They also hypothesize these larger chest tubes are as painful as tragically possible, and the tradition of large chest tubes results in undue suffering.
The answer…remains unsettled. There was no difference observed in their analysis of chest tubes of maximum size versus smaller-than-maxiumum size. But, a 28-32 Fr chest tube is still a pretty darn large tube. I’m not surprised that pain and drainage characteristics were not different. If they really wanted to push the envelope, they ought to look at the truly small tubes – at least, if their goal is clinically relevant pain reduction.
“Does size matter? A prospective analysis of 28–32 versus 36–40 French chest tube size in trauma”
www.ncbi.nlm.nih.gov/pubmed/22327984
The NICE Traffic Light Fails
Teasing out serious infection in children – while minimizing testing and unnecessary interventions – remains a challenge. To this end, the National Institute for Health and Clinical Excellence in the United Kingdom created a “Traffic Light” clinical assessment tool. This tool, which uses colour, activity, respiratory, hydration, and other features to give a low-, intermediate-, or high-risk assessment.
These authors attempted to validate the tool by retrospectively applying it to a prospective registry of over 15,000 febrile children aged less than 5 years. The primary outcome was correctly classifying a serious bacterial infection as intermediate- or high-risk. And the answer: 85.8% sensitivity and 28.5% specificity. Meh.
108 of the 157 missed cases of SBI were urinary tract infections – for which the authors suggest perhaps urinalysis could be added to the NICE traffic light. This would increase sensitivity to 92.1%, but drop specificity to 22.3% – if you agree with the blanket categorization of UTI as SBI.
Regardless, the AUC for SBI was 0.64 without the UA and 0.61 with the UA – not good at all.
“Accuracy of the “traffic light” clinical decision rule for serious bacterial infections in young children with fever: a retrospective cohort study”
www.ncbi.nlm.nih.gov/pubmed/23407730
Informatics for Wrong-Patient Ordering
It seems intuitive – if, perhaps, the electronic health record has an updated problem list, and the EHR knows the typical indication of various medications, then the EHR would be able to perform some cursory checks for concordance. If the orders and the problems are not concordant – then, as these authors propose, perhaps the orders are on the wrong patient?
This study is a retrospective analysis of the authors’ EHR, in which they had previously implemented alerts of this fashion in the interests of identifying problem lists that were not current. However, after data mining their 127,320 alerts over a 6-year period, they noticed 32 orders in which the order was immediately cancelled on one patient and re-ordered on another. They then conclude that their problem list alert also has the beneficial side-effect of catching wrong-patient orders.
A bit of a stretch – but, it’s an interesting application of surveillance intelligence. The good news is, at least, that their problem list intervention is successful (pubmed) – because a 0.25 in 1000 patient alert yield for wrong-patient orders would be abysmal!
“Indication-based prescribing prevents wrong-patient medication errors in computerized provider order entry (CPOE)”
www.ncbi.nlm.nih.gov/pubmed/23396543
JAMA, Integrity, Accessibility, and Social vs. Scientific Peer Review
Yesterday, I posted regarding a JAMA Clinical Evidence series article involving procalcitonin measurement to guide antibiotics stewardship. This is an article I read, raised concerns regarding other negative trials in the same spectrum, and depressingly noted conflict-of-interest with each of the three authors.
Graham Walker, M del Castillo-Hegyi, Javier Benitez and Chris Nickson picked up the blog post, spread it through social media and Twitter, and suggested I write a formal response to JAMA for peer-reviewed publication. My response – I could put time into such a response, but what would JAMA’s motivation be to publish an admission of embarrassing failure of peer-review? And, whatever response they published would be sequestered behind a paywall – while BRAHMS/ThermoFisher continued to happily reprint away their evidence review from JAMA. Therefore, I will write a response – but I will publish it openly here, on the Internet, and the social peer review of my physician colleagues will determine the scope of its dissemination based on its merits.
Again, this JAMA article concerns procalcitonin algorithms to guide antibiotic therapy in respiratory tract infections. This is written by Drs. Schuetz, Briel, and Mueller. They each receive funding from BRAHMS/ThermoFisher for work related to procalcitonin assays (www.procalcitonin.com). The evidence they present is derived from a 2012 Cochrane Review – authored by Schuetz, Mueller, Christ-Crain, et al. The Cochrane Review was funded in part by BRAHMS/ThermoFisher, and eight authors of the review declare financial support from BRAHMS/ThermoFisher.
The Cochrane Review includes fourteen publications examining the utility of procalcitonin-based algorithms to initiate or discontinue antibiotics. Briefly, in alphabetical order, these articles are:
- Boudama 2010 – Authors declare COI with BRAHMS. This is a generally negative study with regards to the utility of procalcitonin. Antibiotic use was reduced, but mortality trends favored standard therapy and the study was underpowered for this difference to reach statistical significance (24% mortality in controls, 30% mortality in procalcitonin-guided at 60 days).
- Briel 2008 – Authors declare COI with BRAHMS. This study is a farce. These ambulatory patients were treated with antibiotics for such “bacterial” conditions as the “common cold”, sinusitis, pharyngitis/tonsilitis, otitis media, and bronchitis.
- Burkhardt 2010 – Authors declare COI with BRAHMS. Yet another ambulatory study randomizing patients with clearly non-bacterial infections.
- Christ-Crain 2004 – Authors declare COI with BRAHMS. Again, most patients received antibiotics unnecessarily via poor clinical judgement, for bronchitis, asthma, and “other”.
- Christ-Crain 2006 – Authors declare COI with BRAHMS. This is a reasonably enrolled study of community-acquired pneumonia patients.
- Hochreiter 2009 – Authors declare COI with BRAHMS. This is an ICU setting enrolling non-respiratory infections along with respiratory infections. These authors pulled out the 47 patients with respiratory infections.
- Kristofferson 2009 – No COI declared. Odd study. The same percentage received antibiotics in each group, and in 42/103 cases randomized to the procalcitonin group, physicians disregarded the procalcitonin-algorithm treatment guidelines. A small reduction in antibiotic duration was observed in the procalcitonin group.
- Long 2009 – No COI declared. Unable to obtain this study from Chinese-language journal.
- Long 2011 – No COI declared. Most patients were afebrile. 97% of the control group received antibiotics for a symptomatic new infiltrate on CXR compared with 84% of the procalcitonin group. 85% of the procalcitonin group had treatment success, compared with 89% of the control group. Again, underpowered to detect a difference with only 81 patients in each group.
- Nobre 2008 – Authors declare COI with BRAHMS. This is, again, an ICU sepsis study – with 30% of the patients included having non-respiratory illness. Only 52 patients enrolled.
- Schroeder 2009 – Authors declare COI with BRAHMS. Another ICU sepsis study with only 27 patients, of which these authors pulled only 8!
- Schuetz 2009 – Authors declare COI with BRAHMS. 70% of patients had CAP, most of which was severe. Criticisms of this study include critique of “usual care” for poor compliance with evidence supporting short-course antibiotic prescriptions, and poor external validity when applied to ambulatory care.
- Stolz 2007 – Authors declare COI with BRAHMS. 208 patients with COPD exacerbations only.
- Stolz 2009 – Authors declare COI with BRAHMS. ICU study of 101 patient with ventilator-associated pneumonia.
So, we have an industry-funded collation of 14 studies – 11 of which involve relevant industry COI. Most studies compare procalcitonin-guided judgement with standard care – and, truly, many of these studies are straw-man comparisons against sub-standard care in which antibiotics are being prescribed inappropriately for indications in which antibiotics have no proven efficacy. We also have three ICU sepsis studies included that discard the diagnoses other than “acute respiratory infection” – resulting in absurdly low sample sizes. As noted yesterday, larger studies in ICU settings including 1,200 patients and 509 patients suggested harms, no substantial benefits, and poor discriminatory function of procalcitonin assays for active infection.
Whether the science eventually favors procalcitonin, improved clinical judgement, or another biological marker, it is a failure of the editors of JAMA to publish such deeply conflicted literature. Furthermore, the traditional publishing system is configured in such a fashion that critiques are muted compared with the original article – to the point where I expect this skeptical essay to reach a far greater audience and have a greater effect on practice patterns via #FOAMed than through the traditional route.